$Ii$alSal$ei$a1#^.JalJel#lSal$.t$.1J""1^dr1$"1S.i{6ld61S"i+".tS"l+.1#"tdbl+"1{L"1dU#ffi# I ".DELLTNG

AriD coNrRoL oF cApAcrrrvE DEroNrzArroN I i

E I DESALTl\ArroN sYSTEM I i+lH+lrltl BY ltr* I

''EDADNAN ALr

I ffi#l ;"_;:,_*-;",," lffi* L^..,A .^-'i:::l::i"T::.'::',:l:::i.,\,F.,^r. I FS. I KING FAHD uNrvERsrTy oF nETRoLEUM & MTNERATs

I H*I

DHAHRAN,SAUDI ARABIA IETEI

#l rnporriot Futrilmentorthe lF

*l J;ffiT;NcE lffi

* | sysrEMS AND coNrRoL ENG*{EERTNG

| ffi

#I MAYzArc

IE#t tFffiiry;agryppl?F,Wfry ,fl,,ffi

MODELLING AND CONTROL OF CAPACITIVE DEIONIZATTON

DESALII\ATION SYSTEM

BY

SYED ADNAN ALI

A Thesis Presented to the

DEANSHIP OF GRADUATE STUDIES

KING FAHD UNIVERSITY OF PETROLEUM & MINERATS

DHAHRAN, SAUDI ARABIA

ln Portiol Fulfillment of the

Requirements for the Degree of

MASTER OF SCIENCEln

SYSTEMS AND CONTROL ENGII{EERING

MAY 24rc

iii

© Syed Adnan Ali

2016

iv

Dedicated To

My Parents

v

ACKNOWLEDGMENTS

All praise is due to Allah alone, the Lord of the Worlds. We constantly seek His help and pray for

guidance. I would like to thank Allah for providing me this opportunity to defend my MS thesis.

The thesis would not have been accomplished without His help. The challenges and difficulties

during the course of my MS study made me strong with a zeal to excel further in the area of my

study.

I would like to thank my parents, Mr. and Mrs. Syed Sajid Ali for trusting in me and providing me

with constant encouragement, prayers and support. I would also like to thank my brother, Mr.

Syed Salman Ali for providing motivation and assistance in numerous ways.

I take immense pleasure in expressing my deep gratitude and appreciation to my thesis advisor

Dr. Moustafa Elshafei for his assistance and professional endeavor that has helped me in

completing my thesis work successfully. My task would have been difficult without his interest

and co-operation and valuable suggestions throughout the tenure of my thesis. He kept track of

my progress and instilled in me a passion for research. I have always admired his passion for

research, wisdom and ingenuity. I remember working with him during weekends. I would like to

thank him wholeheartedly for his valuable time and support.

I would also like to express my gratitude to Dr. Faiz ur Rahman for sharing with me his valuable

real time desalination industry experience, that encouraged me to be actively involved in this area

of active research and serve the community.

I would like to thank the faculty of Systems Engineering department, particularly Dr. Abdul Wahid

A. Al Saif who gave me constant motivation and valuable advices throughout my MS and for his

constant help and support. Thanks to him, I got my Schengen visa.

vi

I also express my gratitude to my committee members, Dr. Faizan Mysorewala and Dr. Sayyid

Anas Vaqar for their valuable advices and feedbacks throughout my work.

The technical staff, Mr. Shoukat Ali, Mr. Jose, Mr. Widodo owe a special mention for their support

and help during lab experiments. I am also grateful to Mr. Khasim Farooqui and Mr. Mervin for

their assistance. I had a great time with them during registration assistance sessions.

I would like to thank all my friends, especially Mr. Ahmed Rehan for his frequent help and support.

I would also like to thank my graduate advisor, Dr. Salih Duffua for his constant encouragement.

Lastly I would like to thank King Fahd University of Petroleum and Minerals for providing me

scholarship to pursue MS in a highly active research environment.

I owe a lot of thanks to everyone who supported me throughout my work.

vii

TABLE OF CONTENTS

ACKNOWLEDGMENTS .......................................................................................................... V

TABLE OF CONTENTS ......................................................................................................... VII

LIST OF TABLES.................................................................................................................... XI

LIST OF FIGURES ................................................................................................................. XII

LIST OF ABBREVIATIONS .................................................................................................... XIV

ABSTRACT ......................................................................................................................... XVI

الرسالة ملخص .................................................................................................................... XVIII

CHAPTER 1 INTRODUCTION ................................................................................................. 1

1.1 Freshwater Scarcity ......................................................................................................................... 1

1.2 Salinity ............................................................................................................................................ 3

1.3 Types of Desalination processes...................................................................................................... 4

1.3.1 Multi-Stage Flash (MSF) Distillation ............................................................................................ 5

1.3.2 Multi-Effect Distillation (MED).................................................................................................... 7

1.3.3 Vapor Compression .................................................................................................................... 8

1.3.4 Reverse Osmosis (RO) ................................................................................................................. 9

1.3.5 Electrodialysis (ED) ................................................................................................................... 13

1.3.6 Electrodialysis Reversal (EDR) ................................................................................................... 17

1.3.7 Capacitive De-ionization (CDI) .................................................................................................. 17

1.3.8 Membrane Capacitive De-ionization (MCDI)............................................................................. 18

1.4 Thesis Motivation and Problem Statement ................................................................................... 19

1.5 Thesis Objectives .......................................................................................................................... 21

viii

1.6 Thesis Outcomes ........................................................................................................................... 22

1.7 Thesis Outline ............................................................................................................................... 22

CHAPTER 2 LITERATURE REVIEW ........................................................................................ 24

2.1 Introduction to Capacitive De-ionization (CDI) .............................................................................. 24

2.2 CDI geometries.............................................................................................................................. 25

2.3 Single pass operation vs Batch mode operation ............................................................................ 27

2.4 Modelling of CDI process .............................................................................................................. 29

2.4.1 Electrostatic Double layer (EDL) models ................................................................................... 29

2.5 Operational modes in CDI and MCDI ............................................................................................. 33

2.5.1 Constant voltage (CV) mode of operation................................................................................. 33

2.5.2 Constant current (CC) mode of operation ................................................................................. 35

2.6 Energy requirements in CDI........................................................................................................... 36

2.7 CDI cost ......................................................................................................................................... 38

2.8 Electrode materials used in CDI ..................................................................................................... 38

2.9 Parameters affecting the CDI process............................................................................................ 39

2.9.1 Surface properties .................................................................................................................... 39

2.9.2 Voltage applied to CDI cell ........................................................................................................ 39

2.9.3 Flow rate .................................................................................................................................. 39

2.9.4 Feed solution concentration ..................................................................................................... 40

2.9.5 Treatment time ........................................................................................................................ 40

2.9.6 Temperature ............................................................................................................................ 40

2.10 Review of CDI ........................................................................................................................... 41

2.11 Companies providing CDI based desalination equipment ......................................................... 49

CHAPTER 3 MATHEMATICAL MODELLING AND ANALYSIS OF A CDI CELL ........................... 51

ix

3.1 Introduction .................................................................................................................................. 51

3.2 Modelling of Capacitive De-ionization (CDI) cell ............................................................................ 53

3.2.1 Modelling equations of a CDI cell ............................................................................................. 55

3.3 Description of the AQUA EWP CDI experimental setup ........................................................... 57

3.4 Modelling of AQUA EWP CDI unit ................................................................................................. 59

3.5 Simulation Results ........................................................................................................................ 60

3.5.1 Simulating the experimental AQUA EWP CDI system using the model ............................... 60

3.5.2 Investigating removal efficiencies and other operational parameters ...................................... 64

3.6 Investigation of CDI batch mode of operation using model ........................................................... 67

3.6.1 CDI model – batch mode .......................................................................................................... 67

3.6.2 Results – CDI batch mode of operation ..................................................................................... 68

3.7 Conclusion .................................................................................................................................... 71

CHAPTER 4 SWITCHED WAVE CAPACITIVE DE-IONIZATION ................................................ 73

4.1 Introduction .................................................................................................................................. 73

4.2 Description of SWCDI system ........................................................................................................ 74

4.3 Switched Wave Capacitive De-ionization (SWCDI) cell .................................................................. 77

4.4 Modelling of SWCDI cell ................................................................................................................ 80

4.5 Optimization algorithm to achieve desired response .................................................................... 85

4.6 Simulation Results ........................................................................................................................ 86

4.6.1 Discussion................................................................................................................................. 89

4.7 Investigation of reflux mode of operation of SWCDI ..................................................................... 90

4.7.1 Description of reflux SWCDI ..................................................................................................... 90

4.7.2 Inlet tank model ....................................................................................................................... 91

4.7.3 Modelling of reflux SWCDI ....................................................................................................... 92

4.7.4 Results – Effect of reflux ratio on permeate concentration ...................................................... 92

x

4.8 Conclusion .................................................................................................................................... 96

CHAPTER 5 CDI EXPERIMENTS ........................................................................................... 97

5.1 Introduction .................................................................................................................................. 97

5.2 Zorflex Knitted FM-50K Activated Carbon cloth (ACC) ................................................................... 98

5.3 Experimental setup of a CDI cell .................................................................................................... 99

5.3.1 Description of the experimental setup ................................................................................... 100

5.3.2 Results and Observation ......................................................................................................... 102

5.4 Experimental setup – SWCDI ....................................................................................................... 103

5.4.1 Observations and Recommendation ...................................................................................... 105

CHAPTER 6 CONCLUSIONS AND FUTURE WORK............................................................... 107

6.1 Conclusions ................................................................................................................................. 107

6.2 Future Work ................................................................................................................................ 109

REFERENCES .................................................................................................................... 110

VITAE ............................................................................................................................... 118

xi

LIST OF TABLES

Table 1.1 Classifying water based on salinity [1] ............................................................. 4

Table 2.1 Energy demands of different desalination technologies [42] ......................... 37

Table 3.1 Operation of AQUA EWP CDI unit ............................................................... 58

Table 3.2 Parameters used in AQUA EWP CDI simulation model ............................... 60

Table 3.3 Comparison of TDS removal efficiency between model and experiment [31]

for different TDS concentrations. ................................................................... 65

Table 3.4 Comparison of effluent TDS concentration of the purified stream between

model and experiment [31] for different flow rates. ....................................... 66

Table 4.1 Switching time and operation in SWCDI cell ................................................ 78

Table 4.2 Sign of 𝑠𝑛𝑖for 𝑉𝑖 and 𝑉�� .................................................................................... 84

Table 4.3 Parameters of SWCDI cell .............................................................................. 87

Table 4.4 Optimized parameters of SWCDI system ....................................................... 87

Table 4.5 Parameters used in simulation of reflux SWCDI system ............................... 93

Table 4.6 Permeate, Inlet and brine concentration for different reflux ratio .................. 95

Table 5.1 Dimension of SWCDI system....................................................................... 103

xii

LIST OF FIGURES

Figure 1.1 Freshwater resources in the world [4] ............................................................. 2

Figure 1.2 Average per capita consumption of water for selected countries [5] .............. 3

Figure 1.3 Types of desalination processes [11] ............................................................... 5

Figure 1.4 Multi-stage flash distillation diagram [12] ...................................................... 5

Figure 1.5 Water desalination using MED [12] ................................................................ 7

Figure 1.6 Vapor compression process [14] ..................................................................... 8

Figure 1.7 Normal Osmosis [15] ...................................................................................... 9

Figure 1.8 Reverse Osmosis [15] .................................................................................... 10

Figure 1.9 Operation of reverse osmosis depicting permeate flow and brine flow [16] 11

Figure 1.10 Conventional flow and Cross flow in reverse osmosis [18] .......................... 12

Figure 1.11 RO cross flow [19] ........................................................................................ 12

Figure 1.12 Filtration spectrum by Osmonics [20] ........................................................... 13

Figure 1.13 Electrodialysis process [21] ........................................................................... 14

Figure 1.14 A typical ED configuration showing movement of ions [12] ....................... 15

Figure 1.15 Capacitive De-ionization cell [1]................................................................... 17

Figure 1.16 Membrane Capacitive De-ionization (MCDI) diagram [1] ........................... 18

Figure 2.1 Geometries used in CDI. (a) Flow by CDI, (b) Flow through CDI, (c)

electrostatic-ion pumping, (d) wire based CDI [29] ...................................... 26

Figure 2.2 CDI experiments design (a) Single pass mode, (b) Batch mode CDI showing

variation of outlet conductivity with time [29] ............................................. 27

Figure 2.3 (a)Helmholtz model, (b) Gouy Chapman model, (c) Gouy-Chapman-Stern

(GCS) model [53] .......................................................................................... 30

Figure 2.4 Models for Electrostatic double layers. (a) GCS model for a single planar

EDL describing the EDL structure, (b) Two porosity model of the electrode

[57], [58] showing macropores and micropores [29] .................................... 32

Figure 2.5 Effluent salt concentration with time for different operating modes viz., (a)

CDI ZVD, (b) MCDI ZVD, (c) CDI RVD, (d) MCDI RVD [29] ................. 34

Figure 2.6 Effluent salt concentration with time for MCDI CV mode of operation and

CC mode of operation [29] ............................................................................ 36

Figure 3.1 CDI cell operation during purification – anions gets adsorbed on the anode

and cations to the cathode ............................................................................. 53

Figure 3.2 Equivalent electric circuit of a CDI cell ........................................................ 54

Figure 3.3 Capacitive de-ionization system schematic diagram [31] ............................. 57

Figure 3.4 Diagram illustrating two CDI cells in series ................................................. 59

Figure 3.5 CDI operational cycle using model- purification and regeneration – showing

effluent concentration with time .................................................................... 62

Figure 3.6 Capacitor voltage profile with time ............................................................... 63

Figure 3.7 Captured charge in coulombs in the CDI cell 2 ............................................ 63

Figure 3.8 Current in CDI cell 2 varying as a function of time (in amperes) ................. 64

xiii

Figure 3.9 CDI Batch mode of operation ........................................................................ 67

Figure 3.10 Effluent concentration at the output of CDI cell and in the inlet tank .......... 69

Figure 3.11 Effluent concentration in the inlet tank at different flow rates q ................... 70

Figure 4.1 Block diagram of switched wave capacitive de-ionization system ............... 74

Figure 4.2 Assembly of CDI cells arrangement in a SWCDI cell .................................. 77

Figure 4.3 Equivalent electrical circuit of 3 CDI cells in series ..................................... 79

Figure 4.4 Equivalent electrical circuit during 𝑇𝑠1 .......................................................... 81

Figure 4.5 Equivalent electrical circuit during 𝑇𝑠2 .......................................................... 83

Figure 4.6 Concentration variation in SWCDI cell ........................................................ 88

Figure 4.7 Voltage across capacitor in SWCDI cell ....................................................... 89

Figure 4.8 Diagram of SWCDI system with reflux ........................................................ 90

Figure 4.9 Variation of permeate concentration with time for 𝛼𝑅 = 1 and 𝛼𝑅 = 10 ... 94

Figure 4.10 Concentration in the inlet tank ...................................................................... 95

Figure 5.1 Diagram illustrating CDI operation ............................................................... 98

Figure 5.2 FM-50K Activated carbon cloth attached to the aluminium sheet ................ 99

Figure 5.3 Illustration of experimental setup of a CDI cell .......................................... 100

Figure 5.4 CDI Experimental setup .............................................................................. 100

Figure 5.5 Instrumentation of CDI experimental prototype ......................................... 101

Figure 5.6 SWCDI experimental setup ......................................................................... 103

Figure 5.7 Aluminium sheets attached to the Acrylic sheet ......................................... 104

Figure 5.8 Activated carbon attached to the aluminium sheets using 3M-9713 Electrical

conducting tape ............................................................................................ 105

Figure 5.9 SWCDI system with connections to switches ............................................. 105

xiv

LIST OF ABBREVIATIONS

AC : Activated Carbon

ACC : Activated Carbon Cloth

CC : Constant Current

CDI : Capacitive De-ionization

CV : Constant Voltage

ED : Electrodialysis

EDL : Electric Double Layer

EDR : Electrodialysis Reversal

EWP : Electronic Water Purifier

GC : Gouy-Chapman

GCS : Gouy-Chapman-Stern

IUPAC : International Union of Pure and Applied Chemistry

MCDI : Membrane Capacitive De-ionization

mD : Modified Donnan

MED : Multi-Effect Distillation

MSF : Multi-Stage Flash Distillation

xv

RO : Reverse Osmosis

RR : Recovery Ratio

RVD : Reverse voltage desorption

SWCDI : Switched wave Capacitive De-ionization

TDS : Total Dissolved Solids

ZVD : Zero voltage desorption

xvi

ABSTRACT

Full Name : [Syed Adnan Ali]

Thesis Title : [Modelling and Control of Capacitive De-ionization Desalination

System]

Major Field : [Systems and Control Engineering]

Date of Degree : [May, 2016]

Freshwater scarcity is one of the most challenging problems facing the world today. Water

is necessary for human beings to maintain good health. Although water is present in

abundance in the form of various sources on earth, only less than one percent of it is

available for drinking purposes. To use abundantly available water resources on earth

requires use of desalination technologies to remove impurities and salts from water to make

it fit for drinking purposes.

Many desalination processes are available based on thermal energy, mechanical energy and

electrical energy or a hybrid of these. Capacitive de-ionization (CDI) is a method of water

desalination based on electrical energy that uses porous electrodes (activated carbon) to

store the ions. In this thesis, we develop a model of capacitive de-ionization using its

equivalent electrical circuit and a dynamical equation to describe how the effluent salt

concentration varies with respect to time in a CDI cell. The model effectiveness is

evaluated by comparing its results with electrosorption experiments from the literature.

The typical CDI operates in two steps; ion adsorption, where we get freshwater out of CDI

cell, and ion desorption that results in discharge of highly concentrated brine. A novel

method of capacitive de-ionization called Switched Wave Capacitive De-ionization

(SWCDI) is proposed and modelled to obtain a continuous flow of freshwater and brine

xvii

with time. Upon successful modelling of the system, an optimization of operational

parameters is done to achieve desired freshwater concentration and high water recovery

ratio.

Lastly, the experimental setup and results of a single CDI cell and SWCDI system is

presented.

xviii

ملخص الرسالة

سيد عدنان علي :الاسم الكامل

نمذجة ومراقبة بالسعة نظام تحلية دي التأين عنوان الرسالة:

أنظمة و هندسة التحكم التخصص:

٢٠١٦مايو :تاريخ الدرجة العلمية

تواجه الناس هذه الأيام. كما لا يخفى أهمية الماء في حياة الإنساااان إن ندرة المياه من أكبر المشااااك و التحديال التي

و للمحافظة على صاحته. بالرمم من توفر مصاادر المياه المتنوعة على ظ ر الأرإ إلا أن الصاالل للشاره من ا لا

تحلية المياه نيالتتعدى نساابته واحد في الما.ة. لتسااتفادة من بمية مصااادر المياه البير عذبة يتمله الأمر إسااتخدام تم

لإزالة الشااوا.ه و الأم ل لتصاابل صااالحة للشااره بعد ذلأ. هناأ أنوال من مختلفة من عمليال تحلية المياه من ا ما

إزالة هو معتمد على الماقة الحرارية أو الميكانيكية أو الك ربا.ية أو كان هجينا بين نوعين أو أكثر من هذه الأشكا .

( واحدة من مرق تحلية المياه و تعتمد على الماقة Capacitive De-ionization CDIالأيونال الساااااااعوية

مستخدمة الإلكترودال المسامية إلكترونال مفعلة( لحفظ الأيونال. في هذه الرسالة نموم بتموير نموذج الك ربا.ية

كيفية تبير تركيز الملل توصاااا لإزالة الأيونال السااااعوية باسااااتخدام دا.رة ك ربا.ية مكاف.ة ل ا و معادلة ديناميكية

المتبمي في خلية إزالة الأيونال الساااااعوية بالنسااااابة للزمن. يتم تمييم كفاءة النموذج بممارنة نتا.جه ما النتا. المعملية

في خموتين: الخموة (CDIتتلخك مريمة ا لمريمة الإمتصاااااك الك ربا.ي التي تمل في الدراسااااال السااااابمة.

(، و الخموة الثانية هي إمتزاز CDIلأيونال حيث أننا نتحص على الماء النمي من خلية ا الأولى هي إمتصاك ا

الأيونااال الناااتجااة من تفريل المحلو الملحي ذو التركيز العااالي. هنااالااأ مريمااة جاادياادة من مرق إزالااة الأيونااال

Switched Wave Capacitive Deionization -السااعوية تساامى إزالة الأيونال السااعوية متحولة الموجة

(SWCDI) قد إقترحل و تم إسااااتنتاج نموذج ل ا للحصااااو على إنسااااياه مسااااتمر للماء النمي و الماء المالل ما )

الزمن. بعد النجال في الحصو على نموذج يتم ضبم و موا.مة باراميترال معام ل( النظام للحصو على النسبة

ى نسااااااااااابااااة إساااااااااااتااااعااااادة لاااالااااماااااء.الاااامااااماااالااااوبااااة ماااان تااااركاااايااااز الااااماااااء الااااناااامااااي و كااااذلااااأ أعاااالاااا

xix

و أخيرا يتم عرإ التجاره المعملية و النتا. ل ذه التجاره في حالة خلية إزالة الأيونال السعوية مفردة و في حالة

إزالة الأيونال السعوية متحولة الموجة مرة أخرى.

1

1 CHAPTER 1

INTRODUCTION

1.1 Freshwater Scarcity

Freshwater scarcity is one of the most challenging problems facing the world today. Water

is necessary for human beings for survival. Around 75 percent of the Earth’s surface

is covered by water out of which 97.5 percent is in the form of oceans and a small

percentage attributes to ice mass, rivers and lakes as shown in Figure 1.1. Even though the

water is present in abundance, only 1 percent of it is available for drinking purposes. Rivers,

lakes and groundwater are the available freshwater sources and occupy a fraction of 0.26

% of total freshwater [1]. Over 1.5 billion people lack ready access to drinking water and

approximately 2.3 billion people suffer water shortages around the globe [2]. As a result of

population growth and increasing quality of lifestyle of the people, it is expected that over

two-thirds of the population will be facing freshwater shortage in the near future to a small

or larger extent [3]. Pollution and exploitation of freshwater sources contribute to the

scarcity of freshwater around the world. Freshwater scarcity is a serious issue. It slows or

stops the economic growth of a country, reduces agricultural output, deteriorates public

health and quality of life.

2

Figure 1.1 Freshwater resources in the world [4]

Although a large amount of water exists on the earth, it cannot be used for drinking and

other purposes mainly because of high salinity and presence of bacteria in it. Desalination

aims to remove salt and other minerals from water to make it suitable for drinking purposes.

An example of natural desalination is the water cycle. The water evaporates from the

surface of the oceans, form clouds and some part precipitates to form snow while other part

produces rain, the freshwater. Both the snow and rain are the purest forms of water. Gulf

countries have very little rainfall per annum and thus it is the challenge of the desalination

technologies to provide freshwater to satisfy the needs of the growing population as well

as for industrial and agricultural purposes. These desalination technologies must be

environmentally friendly causing no pollution, cost effective and energy efficient.

3

Figure 1.2 Average per capita consumption of water for selected countries [5]

Figure 1.2 shows the average daily water use per person for different countries in

decreasing order. The countries differ widely in water usage. The per capita consumption

is around 580 liters in the USA and around 500 liters in Australia. The average per capita

consumption in the Kingdom of Saudi Arabia is equal to 265 liters/day. The average water

consumption in KSA is double the average water consumption of the world [6]. This shows

there is large requirement of freshwater in the Kingdom of Saudi Arabia which is met by

the desalination plants.

1.2 Salinity

The salinity of water is defined based on the total dissolved solids (TDS) present in it. It is

measured in parts per million (ppm) or milligrams per liter (mg/l). For dilute solutions,

both the units are interchangeable, but not for concentrated solutions. TDS comprises

4

inorganic salts (cations- calcium, magnesium, sodium etc. and anions- carbonates, nitrates,

chlorides, sulphates etc.) as well as small amount of organic matter.

Table 1 below shows the salinity levels of different types of water. As can be seen that sea

water consists of maximum TDS level. TDS level of drinking water is specified by WHO

as less than 500 ppm [7] though some countries still consume water up to 1500 ppm TDS

[8].

Table 1.1 Classifying water based on salinity [1]

Type TDS (mg/liter)

Sea water 10,000-45,000

Brackish water 1,500-10,000

Salt water >10000

Freshwater up to 500

Large amounts of sodium in water cause high blood pressure [9] whereas excessive fluoride

in water cause a lot of dental problems [10]. The concentration of some of these ions must

be lowered so as to lower the TDS of water and make it suitable for drinking, domestic and

agricultural purposes. This is achieved using desalination processes.

1.3 Types of Desalination processes

Several desalination processes are used based on utilization of thermal energy, mechanical

energy and electrical energy as shown in Figure 1.3.

5

Figure 1.3 Types of desalination processes [11]

The majority of plants in Gulf region for sea water desalination are based on thermal

energy. The description of some of the thermal energy plants is as follows:

1.3.1 Multi-Stage Flash (MSF) Distillation

Figure 1.4 Multi-stage flash distillation diagram [12]

6

MSF plant consists of three main parts namely, the heat input section, the heat recovery

section, and the heat reject section as shown in Figure 1.4. As the sea water passes through

the tubes, it gains heat. After reaching the first stage, the remaining heat to the sea water is

provided by the steam of the boiler in the brine heater. The steam flows outside the tubes

whereas the seawater flows inside the tubes. As a result, heat exchange process occurs

raising the temperature of sea water. The stages in MSF plants are enclosed and sealed

tightly to create a vacuum. Every stage is maintained at a specific pressure. The pressure

in the stages decreases from 1st stage to the last Nth stage. The sea water after gaining heat

from the brine heater is still in liquid form because it is maintained at high pressure. As the

water enters the orifice, the opening into the 1st stage, because of difference in pressure

(lower pressure as compared to before), there is vigorous flashing. The vapors pass through

demister pads to not allow any brine or impurity along with the vapor and these pure vapors

condenses in contact with the tube bundles (exchange of heat raising the temperature of

sea water passing from Nth stage to 1st stage). The brine stream from 1st stage passes on to

the next stages lower in pressure than the previous stages and similar flashing occurs

resulting in the formation of pure water. The condensed vapors from all the stages are

collected as pure freshwater (approx. 0 ppm TDS) and the brine discharge is discarded.

A majority of desalination plants in Gulf region operates on this technology. MSF plants

are capable of delivering pure water (0 ppm TDS theoretical). These plants have a lot of

potential and many advanced plants based on MSF are in use called Dual purpose plants

that perform the function of generating electricity as well as desalination. These plants are

most widely used in Gulf region.

7

1.3.2 Multi-Effect Distillation (MED)

Figure 1.5 Water desalination using MED [12]

The process of MED plant utilizing horizontal tubes is illustrated in Figure 1.5. The stages

here are called effects. The steam from boiler here is passed inside the tubes rather than

outside as in the case of MSF plants. The sea water is sprinkled on these hot tubes as a

result of which boiling occurs and hot vapors formed are passed on to the next stages or

effect via the tubes. The decrease in pressure of the chambers is the principle used in this

process. The hot vapors enter the next effect wherein again the saline water is sprinkled on

the tubes to generate pure vapors. The hot vapors are condensed and collected as freshwater

whereas the brine is collected separately and discharged into the sea. Here, we require heat

only to the first stage and no subsequent heating of all the stages are required as in MSF

plants. These plants are prone to more scaling problems as the water here is sprinkled on

the tubes.

Water recovery which is defined as the ratio of output freshwater flow to the input

feedwater flow is approximately 10 – 15 % for MSF and MED plants [13].

8

1.3.3 Vapor Compression

Figure 1.6 Vapor compression process [14]

Vapor-compression desalination method is based on the following principle. The process

of distillation begins in the boiling chamber. When both the heating elements are switched

on, the boiling process initiates. The compressor is turned on as the temperature of the

water inside boiling chamber approaches boiling point. During boiling, #2 heating element

is off and #1 element is cycled on and off so as to maintain the boiling at the right

temperature in order to maximize efficiency. The steam produced passes through a baffle

system and enters the compressor. The compressor pressurizes the steam that increases its

temperature and this steam is then passed inside the tubes of heat exchanger. Exchange of

heat occurs from the compressed steam that causes water to boil in the chamber to create

more steam. During the process of giving the latent heat, the steam condenses and a hot

distilled water slightly lower than the boiling point is obtained. During the process, either

#1 or #2 elements are cycled on and off to provide the heat required to keep the system at

optimum temperature and to ensure high efficiency [14].

9

An electric heater, e.g., an initial supply of steam is needed to start the system. Since

compression energy is much less than the thermal energy in the form of latent heat of prime

steam, therefore VC process is more efficient than the MED process. This desalination

process has very low specific energy consumption due to which the overall costs associated

is also on the lower side. The above design of Vapor compression (VC6000) utilizes 0.085

kWh energy to produce one US gallon distilled water [14].

1.3.4 Reverse Osmosis (RO)

The majority of desalination systems relying on mechanical energy are based on the

principle of Reverse Osmosis.

Figure 1.7 Normal Osmosis [15]

Osmosis needs to be understood first in order to understand reverse osmosis [9]. In osmosis

as shown in Figure. 1.7, there is diffusion of a solvent from a region of low solute

concentration to a region of high solute concentration through a semi-permeable membrane

until concentrations on both sides of the membrane are equal. The semi-permeable

membrane in osmosis is permeable to the solvent molecules, but not to the solute, that

10

results in a chemical potential difference across the membrane thus driving the diffusion.

In order to stop the flow of solvent across the membrane, a pressure difference needs to be

applied to the solution side which is known as the osmotic pressure.

Figure 1.8 Reverse Osmosis [15]

In reverse osmosis [9] as shown in Figure. 1.8, a pressure greater than the osmotic pressure

is applied to overturn the normal osmotic flow. Upon applying pressure to the saline

solution, the water passes from more concentrated side to the less concentrated side through

a semi-permeable membrane. The membrane is a thin microporous surface that allows

water to permeate through it while blocking or rejecting impurities, bacteria, pyrogens and

high percentage of inorganic solids. It also rejects polyvalent ions much easier than

monovalent ions.

11

Figure 1.9 Operation of reverse osmosis depicting permeate flow and brine flow [16]

The process cannot go on indeterminately without the removal of impurities. The

impurities and salt present in the saline water end up clogging the membrane that could

reduce the efficiency of the process to produce freshwater. This is referred to as fouling.

Fouled membranes require more pressure to force water through the membrane, thus

causing more energy wastage. As a solution to the above problem, the membranes are

configured so that the feed water coming to the membrane splits into two streams, one part

to be purified as indicated by ‘4’ and the other part to wash away the particles rejected by

the membrane as indicated by ‘5’ in Figure 1.9. There are two types of membranes

available, spiral wound membranes and hollow fine fiber. Each has its own advantages and

disadvantages.

There are two types of flow possible: conventional filtration and cross flow filtration as

shown in Figure. 1.10. The process of conventional filtration is more prone to fouling as

the water flow occurs perpendicular to the membrane under pressure and larger molecules

tend to accumulate on the surface of the membrane. In cross flow, the flow of water occurs

12

tangentially to the membrane surface under pressure. The larger molecules do not

accumulate on the membrane surface but instead move along with the tangential flow of

water [17].

Figure 1.10 Conventional flow and Cross flow in reverse osmosis [18]

The process of Reverse Osmosis depicting cross flow is shown in Figure. 1.11.

Figure 1.11 RO cross flow [19]

Figure. 1.12 below shows the filtration spectrum by Osmonics. It can be clearly seen that

reverse osmosis is capable of removing materials even of ionic sizes up to as small as 0.001

micrometers. The salts and metal ions belong to the ionic range and is removed largely by

the process. The particles in the higher ranges are removed via micro, ultra and nano-

filtration respectively. The main cause of reverse osmosis gaining popularity is because of

13

its low specific energy consumption, 3-8 kWh of electric energy per 𝑚3of freshwater

produced from sea water.

Figure 1.12 Filtration spectrum by Osmonics [20]

There is also a possibility of desalination via electrostatic means that is by supplying

electrical energy. Some of the processes utilizing electrical energy for desalination are

explained below.

1.3.5 Electrodialysis (ED)

Electrodialysis (ED) [12] is a membrane process driven by voltage. ED was introduced

commercially in the 1960s whereas reverse osmosis was introduced commercially in the

1970s. ED is mostly used for brackish water desalination.

14

The basic principle underlining this technology is as follows. Most of the salts dissolved in

water are in ionic form and they allow electric current to pass through them. Some ions are

positively charged (cations) while others are negatively charged (anions). The electrolytic

cell consists of two electrodes connected to a voltage source. The electrodes contain charge.

The ions under the effect of the electric field move towards electrodes having an opposite

charge. The cell consists of ion selective membranes (cation selective membranes that

allow cations to pass through it blocking anions and anion selective membranes that allow

only anions while blocking cations). These membranes are placed alternately in the

electrolytic cell as shown in Figure 1.13 and Figure 1.14.

Figure 1.13 Electrodialysis process [21]

When the saline solution is fed to the cell, the ions electromigrate towards electrode of

opposite polarity. Ion exchange membranes are placed such that it allows certain ion to

15

pass through them and restricts the flow of other type of ions. This process results in the

formation of two compartments, brine compartment where the ions are present in high

concentration and freshwater stream with little or no ions present. Electrodialysis cells can

be arranged in stack configuration as well. The electrodes are usually made of Titanium

(Ti) with Platinum (Pt) coating. During ED operation, oxygen and chlorine gases evolve at

the anode and hydrogen gas at the cathode. There is a continuous supply of freshwater

using this approach as we have separate compartments for both brine as well as freshwater.

Figure 1.14 A typical ED configuration showing movement of ions [12]

Anode and Cathode Reactions

Depending upon pH of the water and potential applied, anodic and cathodic reactions

occurring at each electrode are as shown below.

At cathode,

2e- + 2 H2O → H2 (g) + 2 OH-

16

while at the anode,

H2O → 2 H+ + ½ O2 (g) + 2e-

2 Cl- → Cl2 (g) + 2e-

Hydrogen gas is produced at the cathode and oxygen or chlorine is produced at the anode.

These gases are consequently discharged as the electrode stream effluent from each

compartment of the ED cell. Hydrogen gas evolved can be used for various energy

production applications.

The pumping pressure required in an ED process is only 50-75 psi, all pumps, piping, and

fittings are made of PVC. This method, however, requires periodic cleaning of membranes

by circulating solutions of detergent or other cleaning agents, or by disassembling and

manual cleaning.

[22] mentions that the energy required for electrodialysis can be reduced by 60-70% as the

temperature increases from ambient temperature to 70 ℃. Energy consumption decreases

by 1 % as temperature increases by 0.5 ℃ for temperatures above 21 ℃ [23]. However,

higher temperature damages the membranes and spacers.

Some of the limitations of ED are as below:

1. Polarization Effect: In product water compartments, conductance is very low due

to low TDS. The polarization effect causes the splitting of water molecules to dissociate

into H+ and OH- ions. [24]

2. Electric power required is directly proportional to TDS of saline water. It is

economical only for brackish waters of TDS 3500 ppm [25].

17

1.3.6 Electrodialysis Reversal (EDR)

A modified form of ED is Electrodialysis Reversal. It utilizes the same principle of

electrodialysis. During the operation, the polarity of the electrodes is reversed many times,

so that the ions are attracted in the opposite direction each time the polarity is reversed. At

the same time, flows are also switched so that dilute compartment now becomes the

concentrated compartment and vice versa. This would be beneficial in flushing and

breaking scales and other deposits and also prevent membrane fouling in EDR cell. No

scale control chemicals need to be injected.

1.3.7 Capacitive De-ionization (CDI)

Figure 1.15 Capacitive De-ionization cell [1]

In Capacitive De-ionization (CDI) system, saline water is made to pass between a pair of

electrodes connected to a voltage source. A CDI is a two-step process, the first step is ion

adsorption or charging that results in a pure permeate stream where ions are adsorbed in

porous charged electrodes. As the voltage is applied to the electrodes, due to the presence

of an electrostatic field in between the electrodes, there is a movement of ions such that the

positive ions, cations move to the cathode and anions towards the anode. Two layers of

opposite polarity are formed at the electrode-solution interface which is referred to as

18

electrical double layer (EDL) and ions are stored in these EDL’s. In the next step, voltage

is reversed so that the ions get desorbed and flows out of the CDI cell as a brine stream and

thus cause regeneration of electrodes. CDI differs from ED in the sense that it does not use

membranes. The problems of membrane fouling are not present in CDI and since it utilizes

just electrical energy for desalination, it has the potential to be energy efficient, robust

technology for water desalination. The phenomenon of desalination using CDI is as

illustrated in the Figure 1.15.

1.3.8 Membrane Capacitive De-ionization (MCDI)

The version of CDI with membranes is called MCDI. It is the most promising development

in CDI and has a lot of scope in the future. It utilizes ion exchange membranes (ion-

selective) as shown in Figure 1.16 [1].

Figure 1.16 Membrane Capacitive De-ionization (MCDI) diagram [1]

Ion exchange membranes are placed in front of the electrodes, for example, a cation

exchange membrane is placed in front of the cathode which allows cations to pass through

but stops anions to pass through it. When an electric voltage is applied to the electrodes,

due to the presence of electrostatic field in between the electrodes, the ions in the saline

19

water flowing in the spacer channel gets attracted to opposite polarity electrodes. Anions

travel through the anion exchange membrane and gets attached to pores in the anode

whereas cations pass through the cation exchange membrane and gets attached to pores in

the cathode. At the same time the counter-ions are adsorbed in the respective electrodes,

there is co-ion desorption as well. In CDI, these co-ions enters the spacer channel and

reduces the efficiency of the process. But in MCDI, because of the ion selective

membranes, these co-ions are blocked by the membrane and cannot enter the spacer

channel. This results in more purified water in the spacer channel during the adsorption

step. In the desorption step, as the voltage is reversed, the ions are removed from the

electrode pores due to change in potential and are flushed out of the MCDI cell. The

membranes in MCDI does not allow the ions released in the desorption step to get attracted

to their counter electrodes and a concentrated effluent stream is obtained at the outlet of

the MCDI cell. This is often referred to as flushing.

1.4 Thesis Motivation and Problem Statement

The demand for freshwater is increasing by rate of 64 billion cubic meters a year as a result

of increasing world population [26]. There are various desalination technologies that are

available but it is the need of the hour to develop a desalination technology that is energy

efficient, cost effective, environmentally friendly and capable of meeting the ever-growing

demands for water. Thermal desalination is a high-temperature process and consumes lot

of energy [27], [13]. Also, membrane-based desalination, for example, reverse osmosis is

a high-pressure process and has drawbacks associated with the use of membranes [13],

20

[28]. Capacitive De-ionization is proved to be a low pressure and low-temperature process

[29]–[33]. The ions are removed from water rather than water from the ions. It just requires

a small voltage of 1.2 volts for desalination. Electrode cleaning is achieved by short-

circuiting the electrodes of a CDI cell. No special chemicals are required to clean the

electrodes [34]. Also, during the discharging step of a CDI cell to remove the ions by short-

circuiting the CDI cell, the current released can be used to charge a neighboring CDI cell

or a battery. Thus, there is a possibility of energy recovery in CDI [35]–[37]. The cost of

CDI desalination is low compared to RO for same capacity plants as shown in [38]. It is

potentially more energy efficient for brackish water desalination [39], [29], [40]–[42], [38],

[43]. The energy consumption of CDI is less compared to other desalination technologies

[3], [28], [42]. It is also possible to achieve high water recovery (ratio of output flow rate

to input flow rate) using CDI [44], [45].

Different models have been described in the literature to model CDI based on charge

transfer between the electrodes that describe equilibrium and transport properties in CDI

[29], [46], [47]. Electrostatic double layer theory describes the electrosorption of ions and

storage of charge on the electrodes using the concept of the electric double layer. The

double layer is described using different models in the literature [1], [48], [49]. The

effectiveness of the model depends on the ability of the model to describe ions adsorbed in

the double layer inside the electrodes upon application of voltage. The behaviour of the

double layer is highly nonlinear. The double layer forms a capacitor at the electrode-

solution interface and the solution between the electrodes contributes to solution resistance.

The modelling of CDI is a challenging problem which needs to be investigated.

21

A typical CDI is a two-step process [29], [39], [50]. Freshwater is obtained upon applying

a voltage to the electrodes in a CDI cell and highly concentrated salty water (brine) is

obtained by short-circuiting the electrodes in a CDI cell. Freshwater and brine are obtained

in different time slots unlike the other desalination technologies mentioned previously

wherein freshwater and brine are collected at the same time. A configuration of CDI which

combines the two-time slots to produce freshwater and brine at the same time needs to be

investigated.

1.5 Thesis Objectives

The main objectives of this thesis are following:

1. Literature Review of Electrostatic Desalination Systems.

2. Develop a novel mathematical model of Capacitive De-ionization (CDI)

Desalination system using its electrical equivalent circuit.

3. Evaluate the model by comparing the results with reported results from the

literature.

4. Propose a novel capacitive de-ionization technique called, “Switched Wave

Capacitive De-ionization (SWCDI)” that allows continuous operation of CDI

wherein we obtain freshwater and brine at the same time and model the same using

the developed mathematical model of CDI in this thesis.

5. Optimize the parameters of the modelled desalination system to meet the desired

objectives.

6. Develop an experimental prototype of a single CDI cell and possibly the SWCDI.

22

1.6 Thesis Outcomes

Several inland communities lack access to freshwater because of semi-arid to arid climates.

This climate change increases the frequency of droughts. CDI can prove to be an ideal

desalination technology in an inland remote location for the supply of water for drinking

and agricultural purposes as it consumes less power.

The main outcomes of the thesis are as follows:

1. Derivation of a mathematical model of capacitive de-ionization system that

describes how the effluent concentration varies with respect to time in a CDI cell.

2. Studying the impact of influent feed concentration and flow rates on salt removal

efficiency of CDI.

3. Development of a novel design of CDI called “Switched Wave Capacitive De-

ionization (SWCDI)” and derivation of its mathematical model.

4. Optimization of parameters to achieve desired response.

5. Experimental setup of CDI and SWCDI.

1.7 Thesis Outline

In Chapter 2, a detailed literature review of Capacitive De-ionization is mentioned. It

highlights the role of CDI to be a promising technology for brackish water desalination.

Different operational modes of CDI are described. Various electrostatic double layer

models that describe the CDI process are mentioned. The effect of different parameters on

the CDI system performance is mentioned using results from the literature. Finally,

23

different companies providing CDI equipment commercially and for research are

mentioned.

In Chapter 3, we develop a mathematical model of CDI using its electrical equivalent

circuit and a dynamic equation that describes how the salt concentration varies with respect

to time in a CDI cell. The model effectiveness is evaluated by comparing its results with

electrosorption experimental results of AQUA EWP CDI unit. The effect of different initial

feed concentration and flow rates of water on AQUA EWP CDI unit is studied using the

model. Finally, an investigation on modelling of the batch mode of operation of AQUA

EWP CDI unit is described.

In Chapter 4, a novel approach for continuous operation of Capacitive De-ionization

system is proposed using the concept of switched electric wave which is referred to as

Switched Wave Capacitive De-ionization (SWCDI). The mathematical model of a single

CDI cell is modified to model SWCDI system. The parameters of the system are optimized

to achieve desired freshwater concentration and recovery ratio. Also, an investigation of

reflux mode operation of SWCDI is presented.

In Chapter 5, the experimental setup and results of a single CDI cell and SWCDI system is

presented.

Lastly, Chapter 6 summarizes all the work involved in this research along with future work.

24

2 CHAPTER 2

LITERATURE REVIEW

2.1 Introduction to Capacitive De-ionization (CDI)

Capacitive De-ionization process has attracted interest in the recent past mostly because of

its low energy consumption. The CDI cell consists of porous carbon electrodes pair that

gets charged upon application of electrical potential. Upon applying a voltage to the CDI

cell, the ions present in water under the effect of electrostatic field move towards their

respective counter electrodes, cations moving towards the cathode and anions moving

towards the anode and are stored in the electrostatic double layers formed inside the porous

electrodes. This would result in freshwater flowing through the spacer channel in between

the electrodes. The basis of the process is electrostatic double layer formation inside the

intraparticle pores i.e., inside the carbon electrodes known as micropores. The ions are

immobilized as they are adsorbed into the electrostatic double layers. These electrical

double layers become charged and discharged as the cell voltage is varied and stores and

releases the charge. When the micropores get saturated with adsorbed ions, the voltage is

reduced to zero to release the ions from the electrode pores and thus the electrodes are

regenerated. A stack of electrode pairs can be used to improve the efficiency of the process.

This process removes ions from water rather than removing water from the salt and thus

require less energy for desalination.

25

The change in ion concentration with respect to time is measured to determine desalination

using CDI. Normally for CDI experiments, a single salt solution for example, sodium

chloride (NaCl) is used.

As per the International Union of Pure and Applied Chemistry (IUPAC) [29], pores of size

larger than 50 nm are called macropores, pores between 2 and 50 nm are called mesopores

and pores smaller than 2 nm as micropores, which is different than the pore terminology in

porous electrode theory.

2.2 CDI geometries

The majority of the experimental work on CDI utilizes two porous carbon electrodes that

are placed parallel to each other with a small gap in between them through which the feed

water flows. The porous electrodes are attached to a current collector. This small gap is

referred to as the spacer channel. Different CDI geometries are used for the purpose of

water desalination in the literature [29]. In one such configuration as shown in Figure 2.1-

(a), the voltage is applied between the two porous electrodes and the feed water flows in

the space between the electrodes. This configuration is called flow-by CDI as shown in

Figure 2.1-(a).

26

Figure 2.1 Geometries used in CDI. (a) Flow by CDI, (b) Flow through CDI, (c) electrostatic-ion pumping, (d)

wire based CDI [29]

In another configuration of CDI referred to as Flow through CDI as shown in Figure 2.1-

(b), the water flows perpendicularly through the pores in the electrodes. This allows for

faster charging of a CDI cell as the time for ion adsorption to the pores is reduced. This

architecture contributes to high capacity and low energy cost.

Another geometry used is electrostatic ion pumping as shown in Figure 2.1-(c) wherein a

freshwater stream flows at one end and brine stream flows at the other end. The feed

solution is fed incessantly through an opening in the center of the column. Application of

voltage causes opening of valve 1 at one end of the channel allowing freshwater to come

out of the valve while closing of opposite end valve 2. Applying zero voltage causes valve

1 to close and valve 2 to open to allow the brine to flow outside the system [29].

In another configuration known as wire-CDI, movable carbon electrode wires are

employed for desalination as shown in Figure 2.1-(d) [51]. A CDI cell consists of wires

27

with anode wires being placed closely to cathode wires. These cell pairs are placed in saline

water and applying a potential between cathode and anode results in ions adsorption into

their oppositely charged electrodes. This results in lower salinity of water and the electrode

pair is immersed in another water wherein the cell voltage is reduced to zero to cause ions

release from the electrodes. Thus, in this method two streams are obtained, i.e., freshwater

stream and brine stream. The problem of freshwater mixing with that of the influent feed

stream upon switching the voltage is eliminated in this configuration.

2.3 Single pass operation vs Batch mode operation

Figure 2.2 CDI experiments design (a) Single pass mode, (b) Batch mode CDI showing variation of outlet

conductivity with time [29]

A typical CDI cell consists of two porous electrodes separated by a porous spacer layer. A

CDI stack contains a number of CDI cells. Two types of experiments are usually performed

in CDI as shown in Figure 2.2 [29]. In single pass experiment as shown in Figure 2.2-(a),

28

water enters the CDI cell from a feed tank and the water salinity is measured at the outlet

of the CDI cell. The effluent salt concentration decreases initially as the cell voltage is

applied because of ions getting adsorbed in the electrode pores. But as the electrode pores

become saturated with ions, the concentration increases and reaches influent salt

concentration. The amount of salt removed from the solution is calculated by integrating

the salt effluent concentration with time.

In the batch mode operation of CDI as shown in Figure 2.2-(b), the conductivity of feed

water is measured in the reservoir. Here, the feedwater enters the CDI cell from a feed tank

and the low salt concentration water obtained at the output of the CDI cell is fed back to

the tank to further removal of salts. The effluent salt concentration thereby decreases

steadily and reaches a low constant value. The low constant value corresponds to the

situation where all the electrodes are saturated and no more ions can be adsorbed in their

pores. The electrodes are regenerated by applying zero volts between the electrodes and

allowing feed solution to flow in the CDI cell. The concentration of feed rises as the ions

are desorbed back into the solution and then decreases and eventually reaches influent salt

concentration as all the ions are flushed from the system. The total amount of ions removed

here is calculated by taking the difference in initial salinity and final salinity multiplied by

the total volume of water in the system.

Charge efficiency Λ of the electric double layer is a parameter that describes the efficiency

of a CDI cell. For a monovalent salt (NaCl), charge efficiency [1] is

Λ =Equilibrium salt adsorption

Electrode charge

29

If one salt molecule is removed from the solution for one electron transferred from one

electrode to another, charge efficiency Λ would be unity. However, this is not the actual

case as only 0.6-0.8 salt molecule is removed from the salt solution and charge efficiency

is always less than unity [29].

2.4 Modelling of CDI process

Many attempts have been made to model the CDI process [29]. The understanding of the

operation of Capacitive De-ionization and modelling based on charge transfer is discussed

below.

2.4.1 Electrostatic Double layer (EDL) models

The EDL theory [29] describes charge vs voltage and salt vs voltage characteristics of CDI

cell. EDL approach deals with the transfer of charge from one electrode to another, and its

effect on ion concentration inside porous electrode and drop of voltage across the EDLs.

The model is used to describe electrosorption and storage of charge in the micropores

inside carbon electrodes.

There is formation of electrical double layers at the interface of electrode and solution.

Across this interface, there is charge separation that is the electronic charge of the porous

electrode is compensated locally by the charge in electrolyte filled pores. As a result of

charge compensation, overall, EDL is uncharged. The theory of electrical double layer is

attributed to Helmholtz [52]. Helmholtz model assumes all charge on the surface being

compensated directly by adsorption of countercharge on the surface which implies that for

30

each electron transferred from one electrode to the other, one cation is moved to the cathode

and one anion is moved to the anode to compensate for the respective charges considering

monovalent ions. This would result in removal of one salt molecule from the spacer channel

for one electron transferred from one electrode to other. This represents the ideal situation

of CDI. For such a case, charge efficiency Λ would be unity. According to Helmholtz

model, EDL is treated as a simple capacitor wherein a single layer of ions is adsorbed at

the surface. Thus, EDL is considered a molecular dielectric and the charge is stored

electrostatically. The electric potential decreases linearly with distance from the electrode

surface minimizes at the Outer Helmholtz Plane as shown in Figure 2.3 (a). When the

voltage is less than the electrolyte's decomposition voltage, the charge stored varies linearly

with the voltage.

Figure 2.3 (a)Helmholtz model, (b) Gouy Chapman model, (c) Gouy-Chapman-Stern (GCS) model [53]

31

Helmholtz model did not consider diffusion or mixing of ions in solution, adsorption

possibility onto the surface and the solvent dipole moments and the electrode interactions

and insufficiently describes EDL structure in the porous electrodes.

Later on Gouy & Chapman [54], [55] made improvements in the model by introducing a

diffuse model of EDL in which the electrical potential difference decreases exponentially

away from the surface to the bulk electrolyte as shown in Figure 2.3-(b). They observed

that the capacitance depended on the applied potential and the ionic concentration close to

the surface in the diffuse layer and is not considered constant. The ions are maintained at a

distance from the surface and are spread in a layer close to the surface.

Gouy-Chapman model fails for highly charged double layers. It considers ions as point

charges and assumes that there are no physical limits for the ions to approach the surface.

Stern suggested a model combining Helmholtz model and Gouy-Chapman model known

as Gouy-Chapman-Stern (GCS) model [56]. Some ions stick to the electrode giving an

inner Stern layer (as suggested by Helmholtz) while some form outer diffuse layer (as

suggested by Gouy-Chapman) as also shown in Figure 2.4-(a). The Stern layer does not

contain charge. As can be seen from Figure 2.3-(c), up to Outer Helmholtz Plane (OHP),

the behavior of electric potential similar to that in Helmholtz and beyond that it behaves

like Gouy-Chapman that is, decreases exponentially up to diffuse layer.

32

Figure 2.4 Models for Electrostatic double layers. (a) GCS model for a single planar EDL describing the EDL

structure, (b) Two porosity model of the electrode [57], [58] showing macropores and micropores [29]

The width of the diffuse layer is not exact but ion concentrations decrease gradually with

distance from its surface. Figure 2.4-(b) shows the macropores and micropores in the

porous electrode. The macropores are the pores in contact with the solution whereas

micropores are the pores in the porous carbon electrode.

The characteristic thickness of the electric double layer is given by debye length that varies

with concentration in the CDI cell [29].

Inverse Debye length 𝜅 is given as [29],

𝜅 = √2𝐹2𝑥

𝜀𝑟𝜀0𝑅𝑇

(2.1)

Where 𝜀0 is the permittivity of free space, 𝜀𝑟 the relative permittivity of water (78), R

is the universal gas constant (8.3144621 𝐽 𝐾−1𝑚𝑜𝑙𝑒𝑠−1), F is the Faradays constant (96500

couombs) , 𝑥 is the concentration of salt in bulk solution (macropores) and T is the

temperature in Kelvin (297.15 K).

33

Debye length is given by 𝜆𝐷 =1

𝜅.

When the average pore size in micropores of electrodes is smaller than Debye length, there

is overlapping of EDL and Gouy-Chapman-Stern (GCS) model no longer describes the

case of overlapping EDL’s. Modified-Donnan model [46] is used to describe EDL in such

cases.

2.5 Operational modes in CDI and MCDI

Two modes of operation are possible in CDI and MCDI, viz., constant voltage and constant

current operation which are explained below [29].

2.5.1 Constant voltage (CV) mode of operation

During adsorption in CDI, a constant positive voltage is applied and during desorption step,

a zero voltage is applied (zero voltage desorption - ZVD). However, a reversed voltage can

also be applied during ion-desorption step (reverse voltage desorption - RVD). Moreover,

it is also possible to increase the charge efficiency in CDI by increasing the discharge

voltage during desorption step [59].

The results of effluent salt concentration with time for CDI and MCDI for CV- ZVD-mode

of operation and CV-RVD mode of operation are shown in Figure 2.5. It is reported that

the desalination increases when using RVD mode. In ZVD mode, the counter-ions are

released from electrodes and this counter-ions expulsion occurs until the salt concentration

in the macropores becomes equal to the concentration in the spacer channel or until

complete discharging of micropores.

34

Figure 2.5 Effluent salt concentration with time for different operating modes viz., (a) CDI ZVD, (b) MCDI

ZVD, (c) CDI RVD, (d) MCDI RVD [29]

However, in MCDI CV-RVD mode of operation, counter-ions desorption occurs first from

all the electric double layers in the micropores of the porous carbon followed by desorption

from macropores. The macropore salt concentration approaches value close to zero during

salt desorption which leads to effective electrode cleaning of ions from the electrode. Also

in the subsequent adsorption cycle, ions adsorption and capacity is greatly increased

compared to ZVD. The reverse mode of operation is possible in MCDI because when the

counter-ion reaches the spacer channel in the desorption step, even though it experiences

attraction force from the opposite electrode, the membrane in front of that electrode will

not allow counter-ions to penetrate the membrane and eventually ions are flushed out of

the cell.

The use of reversed voltage in CDI desorption step is debated as it is expected that ions

desorbed from the electrodes will be attracted to the oppositely charged electrodes because

35

of the absence of membranes [3], [46]. However, there is a possibility of ions to leave the

stack via channel between the electrodes without being attracted to the electrode of

opposite polarity. [29] mentions that the salt adsorption capacity per cycle is higher in CDI

–CV-RVD compared to CDI-CV-ZVD. Also, there would be frequency doubling in CDI-

CV-RVD owing to concentration decreasing and increasing twice over a single period as

shown in Figure 1.21.

[29], [60] also report salt adsorption in the CV-RVD mode and CV-ZVD mode for CDI

and MCDI where they proved experimentally that the adsorption of salt in MCDI CV-RVD

> MCDI CV-ZVD > CDI CV-RVD > CDI CV-RVD. The RVD mode for both CDI and

MCDI increases the salt adsorption but decreases charge efficiency [61]. Moreover, if

charge efficiency is considered, MCDI is preferable over CDI, the values of Λ for MCDI

is above 0.9, whereas it is below 0.8 for CDI.

2.5.2 Constant current (CC) mode of operation

In CV mode of operation, it is observed that the effluent salinity decreases during

adsorption as ions are stored in the EDL’s and as the cell becomes fully charged the cell

concentration increases to reach the influent salt concentration. In CC-MCDI mode, it is

possible to obtain a constant effluent salt concentration with time as shown in Figure 2.6

[29]. By applying fixed current between the two electrodes, the flux of ions getting stored

in the EDL is also fixed, so as to obtain desired effluent concentration coming out of the

CDI cell. The effluent concentration level can be tuned to a desired value by changing

electrical current and/or water flow rate as explained in [57].

36

Figure 2.6 Effluent salt concentration with time for MCDI CV mode of operation and CC mode of operation

[29]

In CC-MCDI operation, a constant current is applied to the electrodes until the cell voltage

reaches 1.6 volts after which reversed current is applied for ion desorption until zero cell

voltage is obtained. The CC mode in CDI does not help in obtaining a constant effluent

concentration in adsorption step due to the absence of membranes [29].

2.6 Energy requirements in CDI

The minimum energy required for desalination increases with the concentration of the inlet

feed and recovery [42]. The energy demands for some of the desalination technologies

currently used are shown in Table 2.1 and compared with CDI. From Table 2.1, it is clear

that CDI is competitive to RO in relation to energy requirements where CDI consumes

about 0.1 to 2.03 kWh/m3 of freshwater produced whereas RO system utilizes 2-6

37

𝑘𝑊ℎ/𝑚3[42]. Several works have been reported in the literature that discusses the factors

influencing energy requirements in CDI and MCDI [57],[30].

Table 2.1 Energy demands of different desalination technologies [42]

Desalination method Energy demand in 𝒌𝑾𝒉/𝒎𝟑

MSF 10-58

MED 6-58

RO 2-6

ED 0.4-8.7

CDI 0.1-2.03

[28] presents the comparison of MCDI and RO energy demand at different salinities of

NaCl ranging from 10 mM to 80 mM (1 mM = 1 𝑚𝑜𝑙/𝑚3) in order to obtain 5 mM

freshwater. The operation of the equipment for MCDI is using constant current (CC-RCD)

to obtain a water recovery of 50 %. Results demonstrate that there is a crossover point at a

salinity of 40 mM below which MCDI consumes less energy compared to RO. There is a

hope that this crossover point will be shifted to write with the ongoing researches in the

field that will prove the capability of CDI to be a competitive desalination technology for

brackish water desalination.

38

2.7 CDI cost

A direct cost comparison of CDI vs RO was demonstrated in [62]. They compared costs

for brackish water desalination plants of small capacity (3785 𝑚3/day), and claim that CDI

gives a cost of 0.11 $/𝑚3 whereas RO gives a cost of 0.35 $/𝑚3.

2.8 Electrode materials used in CDI

The majority of efforts to improve the CDI technology is done by the development and

synthesis of electrode materials that can improve the desalination performance. Ions are

stored inside the pores of electrodes. Some pores inside carbon do not contribute to ion

adsorption because it is closed from all the sides or is leaky. The electrodes are chosen on

the basis of its cost, ability to tune the pores, surface area of the electrodes, low scalability.

Activated carbons are the most widely used owing to its lower cost, high specific surface

area (1000-3500 𝑚2/gram) and availability [50], [63]. Also other materials have also been

used for CDI and MCDI including carbon aerogels, carbide derived carbons (CDCs), CNTs

and CNFs, mesoporous carbons, graphene etc [32], [64]–[68].

39

2.9 Parameters affecting the CDI process

The main parameters affecting CDI process are explained below:

2.9.1 Surface properties

The presence of surface functional groups including carbonyl, phenol, carboxyl etc.,

impacts the desalination using CDI by enhancing surface wettability and improving the

efficiency of surface utilization [69].

2.9.2 Voltage applied to CDI cell

The high voltage applied to the electrodes increases the thickness of the EDL which results

in strong adsorption of ions. However, if the voltage applied to CDI cell exceeds 1.23 V, it

would result in electrolysis (decomposition) of water. Electrolysis results in a change of

pH in solution and oxidation of carbon electrodes (oxidation of chlorine at anodes [70])

and negatively affects the CDI process [65]. It is also reported that below 1 volts, the pH

increases quickly as potential is increased due to the reduction of dissolved oxygen. It is

concluded by the authors in [71] that the salt removal increases with increasing voltage in

the range of 0.8-1.5 volts but increasing voltage also increases the energy consumption.

2.9.3 Flow rate

[30] showed that upon increasing flow rate from 1 to 4.5 l/min, there is approximately 28%

drop in electrosorption removal efficiency. Electrosorption removal efficiency is defined

as the ratio of the difference in initial and final concentration to the initial feed

concentration. As the flow rate of the inlet feedwater increases, the time for which the ions

40

reside in the CDI cell decreases as a result of which ions do not get sufficient time to

electromigrate to the electrodes that results in low electrosorption efficiency. However,

increasing flow rate results in low energy consumption [72], [73]. Thus, the optimal flow

rate must be governed by both electrosorption removal efficiency and energy consumption.

2.9.4 Feed solution concentration

As the inlet feedwater concentration increases, solution resistance decreases and adsorption

velocity increases which in turn causes the capacitance and salt removal to increase [69].

The salt removal increases as more ions are adsorbed onto the electrodes but the adsorption

ceases when the electrodes reach their saturation limit. [30] showed approximately 41%

drop in electrosorption efficiency when the inlet feed concentration was increased from

500 ppm to 3500 ppm.

2.9.5 Treatment time

As treatment time increases, the ionic concentration decreases at first as ions adsorb onto

the electrodes. However, after some time, the concentration of the effluent begins to rise

as the salt adsorption limit is reached in the electrodes. The adsorption slows down until a

dynamic equilibrium is reached [69].

2.9.6 Temperature

The temperature has a negative effect on salt removal in CDI process. [30] showed 11.2 %

drop in electrosorption efficiency when the temperature of the solution was raised from 20

to 50 degrees Celsius. This is attributed to decrease in adsorption capacity of activated

carbon electrodes with temperature. It is also possible for ions to escape from metal surface

to the solution as the temperature of the feed solution is increased.

41

Other factors including the spacing of the electrodes, number of CDI electrode pairs used,

the configuration of CDI units affects the efficiency of the CDI process. As the distance

between the electrodes is small, adsorption would be stronger and less treatment time is

required. High salt removal efficiency can be obtained by using more pairs of electrode in

CDI.

2.10 Review of CDI

The breakthrough in the CDI technology came from the work of Farmer et al in 1995. It

reported the experimental results of CDI apparatus that was developed at Lawrence

Livermore National Laboratories [64], [74], [75]. Since then the process came to be known

as Capacitive De-ionization (CDI).

[3] explains constant current operational mode for MCDI. MCDI consumes much less

energy per mole of salt removed as compared to CDI. For MCDI, with 1-200 mM ionic

strength feed solution, an energy consumption of approximately 22 kT is reported for one

ion removed where k is the Boltzmann constant and T is the temperature in Kelvin. It shows

that the energy consumption in MCDI does not depend on salt concentration while for CDI

it is directly proportional to the inlet feed solution concentration.

[58] describes MCDI transport model that not only considers double layers formed inside

micropores, but also includes the role of the macropores for ion movements to the

electrode. The spacer channel is considered to be divided into stirred volumes placed in

series and the transport of ions inside spacer channel and from spacer channel to the

42

electrodes is considered by making some assumptions. The experimental setup consists of

8 parallel CDI cells with feed concentration of water equal to 20 mM. The authors

compared the adsorption/desorption cycles of MCDI both experimentally and theoretically

for zero voltage desorption (0-MCDI) and for reverse voltage desorption (r-MCDI), and

also compared with results for CDI. EDL-structure is described by using a modified

Donnan model to describe the salt adsorption and charge in the CDI stack. The authors

based on experimental results claim that the salt removal for 0-MCDI is 20% more than in

CDI, and r-MCDI removes further 20% compared to 0-MCDI making r-MCDI 40% more

effective than CDI. For CDI and 0-MCDI, the variation of charge with time was almost

similar whereas for r-MCDI, charge increased by 20%. The authors suggested the use of

validated MCDI model for further design and process optimization and attributed a little

deviancy due to assuming not ideally permselective membranes (may also allow some co-

ions) and also due to assuming constant chemical attraction term (attraction of ions when

no voltage is applied) in Donnan theory.

[62] presented the development of prototype industrial type CDI module and evaluated its

performance for brackish water desalination to compare its potential to compete with

Reverse Osmosis. The main aim was to convert laboratory unit at LLNL to industrial

prototype unit. The electrodes were subjected to certain quality control tests to ensure

certain minimum standards based on surface area, bulk resistivity, specific capacitance etc.

The energy consumption for desalination of coal bed methane (CBME) produced water

was 2.25 kWh to treat 1000 gallons of produced brackish water. The energy consumption

and various costs of Reverse Osmosis and CDI prototype unit were compared and it was

concluded that CDI technology could replace membrane technologies in the near future.

43

Energy efficiency of CDI plants was more as compared to Electrodialysis Reversal (EDR)

technology plants. For seawater desalination, it was emphasized that the capacitive de-

ionization module production costs need to be reduced in order to compete with reverse

osmosis. It is further mentioned that by utilizing energy recovery processes, industrial units

can approach the laboratory energy consumption.

[76] reports an experimental approach of continuous electrodeionization through

electrostatic shielding. The permselective membranes in the cell were replaced by

electrostatically shielded zones. The electrostatically shielding of compartment causes the

electric field to be zero inside the compartment. The authors claim using this configuration

to regenerate ion exchange resins as well as for producing deionized water and desalinating

brackish (0.02 M NaCl) or sea water. Co-ions are carried by counter-ions migrating in the

shielded compartments. No drawbacks of concentration polarization existed in this process.

[77] presents a novel process for water desalination and electrodeionization by means of

electrostatic shielding zones. These zones are used as ionic current sinks since the electric

field becomes zero inside the zone when an external field is applied to the overall cell. Ions

get concentrated in the sink whereas the other compartments get depleted of ions. 0.03 M

NaCl solution is desalinated to obtain potable water. Even industrial effluents consisting

of heavy metal ions are purified to obtain low concentration pure water. It also specifies

regeneration of ion exchange resins to get high purity deionized water.

[78] modeled the CDI adsorption-desorption processes of CDI to obtain conductivity of

freshwater with respect to time for a given initial feedwater concentration of saline water.

The authors claim that to improve desalination performance and to obtain less concentrated

44

freshwater, number of CDI cells has to be increased. A stack of multiple CDI cells could

be used to achieve required desalination performance.

[79] presented the transient response of the CDI cell operating in a batch mode wherein a

particular volume feed water is fed back to the container incessantly until the system

reaches steady state. The model for the batch system is developed by considering single

pass CDI operation and the mixing phenomena that occur in the recycling tank. The model

was validated using experimental results.

[80] presented experimental data of a pilot plant for outlet salt concentration and current in

MCDI stack as a function of flow rate, inlet ionic strength and time. The position of anion

exchange and cation exchange membranes are reversed in each of the cells because one

electrode acts as a cathode of one cell while acting as the anode for other. The CDI

configuration utilizes “stop-flow” operation mode that helps in concentrating all the stored

salt in 25-40% volume of dilute product resulting in a small and highly concentrated brine

stream. A theoretical model based on GCS theory is described for MCDI which describes

the effluent ion concentration and electric current varying with time and storage of ions in

the EDL both during the adsorption and desorption. The modelling of ion exchange

membranes is done by considering stagnant diffusion layers placed in front of the

membranes. Experimental results showed that with increase of the flow rate and ionic

concentration, the effluent ionic strength increases which would also cause the current to

decrease with time because at lower flow rates the average concentration in the cell is low

leading to increase in resistance for ion transport which ultimately leads to low salt removal

rate and less current. The agreement between the theory and experimental results is not

45

perfect but still satisfying. The authors suggested to improve MCDI model by introducing

additional resistance located within the spacer channel and electrode.

[81] mentions the use of carbide-derived carbon (CDC) as electrodes for CDI by comparing

it with commercially available activated carbons (AC) and demonstrated that CDC

electrodes have a high electrosorption capacity and charge capacity (approximately 28-

44% increase for salt adsorption and 23-34% increase for charge) when voltages in the

levels of 1.2-1.4 V is applied to the system with a particular feedwater concentration and

feedwater flow rate. The authors state that materials having high charge storage capacity

can be highly suitable for CDI application. Modelling of the system is performed using

mD-model.

[82] presents a method to optimize the water recovery in MCDI and the electrosorption

rate by demonstrating how the variation of each parameter individually affect the CDI

operation and performance such as adsorption and desorption time, feed salt concentration,

current and voltage applied to the CDI cell.

[83] demonstrates the use of electret technology to desalination. The authors tried to

reproduce the patent results of Douglas MacGregor who claimed the use of electrets for

water desalination based on electrets being permanently polarized materials. The

permanent electrostatic field attracts counter-ions to the charged surface to produce a

purified effluent in between the electrets. The method was evaluated both theoretically and

experimentally to investigate practical aspects of the technology. The authors reported that

the design is satisfactory only when a dilute feed water is used and the system must have a

46

very high surface charge density to achieve desalination. The authors consider using

electrets for desalination as viable and recommend further development.

The modelling of the electrostatic double layer is described using equilibrium GCS model

in [47]. A description of transfer of ions from bulk solution to the electrodes via a mass

transfer layer is explained. The model very well predicts the experimental datasets of

current and salt concentration in the effluent water with time. A new term, differential

charge efficiency is described based on GCS-model, which is the ratio of rate of salt

removal relative to the electronic current. It describes the ions removed from the solution

for each electron transferred from one electrode to another. During modelling, the authors

assumed two electrodes to be completely non-Faradaic, that is, flow of current due to

electrochemical reactions from electrodes into the solution or vice versa is zero. An

approximate approach based on mass-transfer boundary layer placed in front of electrodes

is used for the charge transport from bulk solution to electrodes. The charge efficiency

obtained using theoretical model was slightly high (0.81) compared to experimental results

(0.77). The authors suggested this drawback due to the fact that pores inside carbon must

be too minute to for ions to enter and remove ions. Also because of overlapping of double

layers, the pores inside carbon leads to lower salt adsorption. The author considers

development of the model to encompass all the drawbacks and deviation from experimental

results.

[84] calculates charge efficiency (ratio of salt adsorption and electrode charge) as a

function of voltage and salt concentration experimentally, and this data is used to describe

the structure of double layer inside electrode and measure the effective area for ion

adsorption. First, the authors calculate charge efficiency by varying cell voltage from 0 to

47

1.4 Volts and ionic strength in the range of 5mM and 20mM based on the data for charge

vs cell potential and salt adsorption vs cell potential. The authors used GCS double layer

model to accurately describe data for charge efficiency by considering stern layer capacity,

𝐶𝑠𝑡 as an adjustable parameter obtained from experimental data set of charge efficiency

varying with voltage and ionic strength. An optimized value of 𝐶𝑠𝑡 is used in GCS model

and the same model is used to fit the full data sets for charge and salt adsorption which in

turn calculate the effective surface area per gram of electrode, 𝑎𝑚. These two optimized

values 𝐶𝑠𝑡 and 𝑎𝑚 of porous electrodes are used as input parameters in electrokinetic

process model and compared with data for ion concentration of the effluent and current

based on different voltages applied to the cell, different flow rates and different inlet ionic

concentrations. The theoretical results matches the experimental results quite effectively.

The double layer is assumed to be at equilibrium with the solution in the bulk. This paper

improves the work of [47] by describing both ion removal and ion release by the same set

of parameters in GCS model.

[59] showed that charge efficiency in CDI can be increased by carrying out desorption step

at a voltage of 𝑉𝑑𝑖𝑠𝑐ℎ = 0.3 volts instead of zero volt desorption in CV mode of operation

of CDI. This caused small reduction of salt adsorption per cycle,the energy per ion removed

decreased by about 14 % in case when there is no energy recovery and led to 30 % decrease

in case of perfect energy recovery. For CC operational mode with 0.41 ampere current

applied to CDI stack consisting of four CDI cells, the authors claim that the constant

effluent concentration level reached swiftly with increased discharge voltage. The

improved mD model [48] describes data set of equilibrium salt adsorption and charge

48

perfectly. The dynamics of ion transfer and charge storage both for CV and CC operation

was illustrated by making use of porous electrode transport theory.

[85]–[87] illustrates continuous mode of desalination using CDI wherein freshwater and

brine are produced at the same time. In this configuration of CDI, slurry electrodes are used

that flow between the porous separator (ion selective membranes are also used as porous

separator) and the electrode. The ions electromigrate to the slurry electrodes and are

flushed as brine whereas freshwater is obtained from between the porous separators.

Research is done to make CDI more efficient and cheaper technology by using the concept

of energy recovery [36], [35]. During the ion desorption step, a significant amount of

energy is lost. This energy can be used to charge a neighbouring CDI cell or supercapacitor

or battery.

[36] uses a buck-boost DC/DC converter in order to charge a neighbouring supercapacitor

using ion desorption of a CDI cell. They claim to obtain a high energy recovery of 51%

using buck-boost converters. The simulation and experimental results are obtained with

minor deviations.

Many attempts have been made in the literature to model the CDI process based on charge

transfer between electrodes utilizing Gouy-Chapman-Stern [47], [84] and Modified

Donnan models [3], [60], [32], [48], [88] and improved modified Donnan models [59], [49]

to describe the electrostatic double layer in the electrodes. [48] describes the transport

model for membrane capacitive de-ionization where it uses modified Donnan model to

describe the salt adsorption and charge in a cell. [59] uses improved modified Donnan

model (i-mD) to describe data set of equilibrium salt adsorption and charge and presents

49

dynamics of ion transfer and charge using porous electrode transport theory. The modelling

of batch mode operation of CDI is shown in [29], [89].

Electrical circuit designs have also been used to characterize a CDI cell [36], [35], [90],

[91]. The most commonly used CDI equivalent circuit is the Randle’s circuit [36], [35],

[90], which takes into account the solution resistance, double layer capacity and leakage

resistance. This model was used to compute energy recovery in a CDI cell utilizing

different converters to recover energy in different ways. [90] compared the electrical

response of CDI cells with the experimental data using Randle’s equivalent circuit. The

model response of CDI cell showed good agreement with experimental data, except for the

prediction of solution resistance, which was overestimated.

Patents by MacGregor mention an apparatus for separating ions from liquids to produce

freshwater and brine using permanently polarized materials electrets wherein the static

electric fields are created in between the electrets that allow the ions of opposite charge to

move towards the counter polarized materials [92], [93].

2.11 Companies providing CDI based desalination equipment

Many companies manufacture CDI equipment for research purposes as well as for

commercial purposes. AQUA EWP [94] has manufactured a commercial equipment for

brackish water as well as for sea water desalination and these equipments can be purchased.

However other companies like PLIMMER CDI [95] also has similar principle based

50

equipment. The Cap DI unit is available for research by the company Voltea BV [96].

Atlantis technologies [97] have used supercapacitors for water desalination.

51

3 CHAPTER 3

MATHEMATICAL MODELLING AND ANALYSIS OF A

CDI CELL

3.1 Introduction

This chapter proposes a novel mathematical modelling of capacitive de-ionization (CDI)

system with activated carbon electrodes. A mathematical model of CDI is presented based

on its electrical equivalent circuit and a dynamic equation that describes how the effluent

salt concentration varies with respect to time in a CDI cell. The model is developed by

taking into consideration the specific surface area of the electrodes, capacitance of the

electrodes, double layer thickness, initial feed concentration, the flow rate of feed water,

volume of the CDI cell, voltage applied to the CDI cell and solution resistance in the CDI

cell. The model predicts the amount of charge captured by the CDI cell and the current in

the CDI cell, both as a function of time. The effectiveness of the model is evaluated by

comparing its results with the reported experimental results of AQUA EWP CDI unit [30],

[31]. The model also determines the performance of the AQUA EWP CDI cell to different

operational parameters i.e., the feed TDS concentration and flow rates. The model results

are in good agreement with the experimental results. The batch mode operation of CDI is

investigated using the derived model.

52

The chapter is organized as follows. First, a mathematical model of capacitive de-

ionization with activated carbon electrodes is derived using an electrical equivalent circuit

of CDI in Section 3.2 that takes into effect the specific surface area of the electrodes,

capacitance of the electrodes, double layer thickness, initial feed concentration, flow rate

of feed water, volume of the CDI cell, voltage applied and solution resistance in the CDI

cell. Next, a CDI unit from AQUA EWP using activated carbon electrodes is explained in

Section 3.3. The CDI model is extended to describe AQUA EWP CDI unit consisting of

two CDI cells in series in Section 3.4. Finally, the model results are compared with the

reported results of electrosorption [30], [31] conducted on a CDI unit from AQUA EWP

in Section 3.5. The model is used to determine the efficiency of the system in treating

brackish water at different feed TDS concentrations and also to evaluate the effect of flow

rate on treatment efficiency. The model is further used to investigate the batch mode

operation of AQUA EWP CDI unit in Section 3.6.

53

3.2 Modelling of Capacitive De-ionization (CDI) cell

++ +

- --

Le

u1

-u1

deq q

oin

dc

Figure 3.1 CDI cell operation during purification – anions gets adsorbed on the anode and cations to the cathode

The CDI cell as shown in Figure 3.1 is a conduit of rectangular cross section of width 𝑊𝑒,

length 𝐿𝑒 and height 𝑑𝑒. CDI cell consists of electrodes (activated carbon) of thickness 𝑑𝑐,

and each electrode of a CDI cell is attached to a current collector. The activated carbon

electrode layer is merged and fully in contact with the salty water. As shown in figure 2,

the top metal collector is connected to a dc voltage signal of +𝑢1 and bottom metal

collector to a dc voltage signal of −𝑢1 during purification that causes the anions to move

towards anode and cations to move towards cathode. During regeneration, the CDI cell is

powered off with zero voltage or a voltage of opposite polarity is applied to the current

collectors. The purification and the regeneration step completes the CDI cycle. The water

flows at a flow rate of 𝑞𝑖𝑛 to the CDI cell and the flow rate at the exit 𝑞𝑜 can be controlled

using a valve.

54

The capacitance associated with each electrode is given by

𝐶𝑒 =𝑘𝑐𝜀0(𝐿𝑒𝑊𝑒)

𝑡𝑐𝑠 (3.2)

where 𝑘𝑐 depends on the effective area ratio of the activated carbon and its wet permittivity,

𝑡𝑐𝑠 is the thickness of the double layer which is of the order of nanometers. The electrical

double layer thickness is often approximated as Debye length using Gouy Chapman model

[29].

𝑘𝑐 = 𝑘𝑠𝜌𝑐𝑡𝑐; where 𝑘𝑠 is the specific surface area of the Activated carbon which is

typically around 800-2500 𝑚2/𝑔𝑟𝑎𝑚, 𝜌𝑐 is the density of AC, 𝑡𝑐 is the thickness of

Activated carbon electrode.

The equivalent circuit of a CDI cell with two electrodes is shown in Figure 3.2.

-U1

U1

C1

C2

IRw

Figure 3.2 Equivalent electric circuit of a CDI cell

The top activated carbon electrode capacity is represented by capacitor 𝐶1 which is

positively charged and bottom electrode capacity is represented by capacitor 𝐶2 which is

negatively charged. The capacitance associated with the two electrodes is equal to maintain

charge balance in solution i.e., 𝐶1=𝐶2.

55

𝑅𝑤 is the electrical resistance of the salty water between the two opposite electrodes and I

is the current flowing in the CDI cell through 𝑅𝑤 resistance.

The resistance of water is given by the formula

𝑅𝑤 =

𝜌𝑒𝑑𝑒

𝐴𝑐

(3.3)

where 𝐴𝑐 is the physical area of the activated carbon in 𝑚2

𝑑𝑒is the distance between the activated carbon electrodes

𝜌𝑒 is the resistivity of the water.

The resistivity of water depends on temperature and concentration and for saline water is

given by the approximation [98], [99],

𝜌𝑒 ≅

5

𝑥(1 + 𝛼𝑇(𝑇 − 25))

(3.4)

Where 𝑥 is the concentration of salt in kg/𝑚3 (or grams/l); and 𝛼𝑇 = 0.022/℃

3.2.1 Modelling equations of a CDI cell

Let 𝑉𝑐 be the voltage across the capacitor, and 𝐼 be the current flowing throw 𝑅𝑤 resistance

and 𝑞 be the flow rate of feed water to the CDI cell.

Let the initial feed concentration to the CDI cell be denoted by 𝑥𝑓 and V be the volume of

the CDI cell.

From the equivalent electrical circuit of CDI cell as shown in Figure 3.2, we write the

following salt balance equation,

56

𝑉

𝑑𝑥

𝑑𝑡= 𝑞(𝑥𝑓 − 𝑥) − 𝐾𝐼

(3.5)

where 𝐾 = 𝑀𝑤𝑠/𝐹, where 𝑀𝑤𝑠 is the molecular weight of NaCl and F is Faradays number

(F=96500 Coulombs).

The current is related to the voltage across the capacitor, 𝑉𝑐 by the formula,

𝐼 = 𝐶𝑒

𝑑𝑉𝑐

𝑑𝑡=

2𝑉𝑖𝑛 − 2𝑉𝑐

𝑅𝑤𝑐

(3.6)

The equations (3.4) and (3.5) can be written as,

𝑑𝑥

𝑑𝑡=

𝑞

𝑉(𝑥𝑓 − 𝑥) −

𝐾

𝑉

(2𝑉𝑖𝑛 − 2𝑉𝑐)

𝑅𝑤𝑐

(3.7)

𝑑𝑉𝑐

𝑑𝑡=

2𝑉𝑖𝑛 − 2𝑉𝑐

𝐶𝑒𝑅𝑤𝑐

(3.8)

Equations (3.7) and (3.8) are the state equation of a CDI cell that describes the relationship

between CDI cell salt concentration and voltage across the capacitor with respect to time.

57

3.3 Description of the AQUA EWP CDI experimental setup

The block diagram of capacitive de-ionization system for desalination is as shown in Figure

3.3.

Figure 3.3 Capacitive de-ionization system schematic diagram [31]

The feed solution used in the experiment [30], [31] is Sodium chloride (NaCl) of desired

concentration. The salty water (NaCl) is first collected in an inlet tank. A pump takes the

salty water from the inlet tank and passes through a sand filter for pre-treatment and water

flows at a flow rate of 𝑞 to the CDI cells. Two CDI cells were used in the experiment. A

potential of 1.5 V DC is applied per CDI cell so that electrolysis does not occur. There are

3 valves in the system viz., influent solenoid valve SV0, outlet solenoid valves SV1 and

SV2. The CDI operational cycle is of 2.5 mins and occurs in two main steps: the

purification mode step (90 seconds) - wherein the inlet valve SV0 and the outlet valve SV1

are opened and a voltage of 1.5 VDC is applied to CDI cells to purify the feed solution and

the regeneration mode that occurs in two steps – in the first 30 seconds, the outlet valve

SV1 and the inlet valve SV0 are closed and the power supply is off that results in

discharging the ionic charge on the electrodes, followed by another 30 s when the outlet

valve (SV2) and the inlet valve (SV0) are opened and the voltage is applied to CDI cells

58

with a opposite polarity of 1.5 VDC that results in removal of ions from the electrodes

along with flushing of highly concentrated brine [31]. The operation of the system is as

illustrated below in the Table 3.1.

Table 3.1 Operation of AQUA EWP CDI unit

Operational cycle

(Time - 2.5 minutes =

150 seconds)

Description Time

(sec)

Feed

pump and

SV0

SV1 SV2

1. Purification Application of 1.5 volts 90 ON ON OFF

2. Regeneration

– 2 steps

1. Discharging the

ionic charge on the

electrodes, no flushing

done

2. Flush mode with

opposite polarity

(negative voltage -1.5

volts)

1. 30

2. 30

1. OFF

2. ON

OFF 1. OFF

2. ON

The recovery ratio of this process is 75 % as it includes 90 seconds of product flow during

purification versus 120 seconds of total flow. The AQUA EWP CDI unit consisted of

activated carbon electrodes with a single CDI cell containing 1354 grams of activated

carbon. The specific area of activated carbon electrodes used was 800 m2/gram. Each cell

assembly contained 200 sheets of activated carbon (100 cathodes and 100 anodes) with

dimensions of 158 mm × 174 mm × 0.3 mm. These electrodes were connected to the two

sides of the DC power supply.

59

3.4 Modelling of AQUA EWP CDI unit

The experiments in the paper [30], [31] consisted of two CDI cells in their setup as shown

in Figure 3.4. In this configuration, the feed water with a concentration of 𝑥𝑓 and a flow

rate of 𝑞 flows through the CDI cell 1. During purification mode of 90 seconds, the water

is purified in the CDI cell 1 and attains the concentration 𝑥1 at the exit of cell. The water

is passed on to CDI cell 2 for further purification and attains a concentration 𝑥2 at the outlet

of CDI cell 2. During regeneration mode, the ions are flushed back into the feed water and

the concentration at the outlet of CDI cells 1 and 2 increases with time. The application of

voltage to both the CDI cells is simultaneous.

u1

-u1

q, xf

u1

-u1

q, x2

q, x1

Saline Water from pump

x1 x2

CDI cell 1 CDI cell 2

Figure 3.4 Diagram illustrating two CDI cells in series

The state equations for CDI cells in Figure 3.4 can be derived using equations (3.6) and

(3.7) as follows that describes the concentration varying in the CDI cells and the voltage

across capacitances with respect to time. The subscripts 1, 2 denote parameters of CDI cell

1 and CDI cell 2 respectively.

60

𝑑𝑥1

𝑑𝑡=

𝑞

𝑉(𝑥𝑓 − 𝑥1) −

𝐾

𝑉

(2𝑉𝑖𝑛 − 2𝑉𝑐1)

𝑅𝑤𝑐1

(3.9)

𝑑𝑉𝑐1

𝑑𝑡=

2𝑉𝑖𝑛 − 2𝑉𝑐1

𝐶𝑒𝑅𝑤𝑐1

(3.10)

𝑑𝑥2

𝑑𝑡=

𝑞

𝑉(𝑥1 − 𝑥2) −

𝐾

𝑉

(2𝑉𝑖𝑛 − 2𝑉𝑐2)

𝑅𝑤𝑐2

(3.11)

𝑑𝑉𝑐2

𝑑𝑡=

2𝑉𝑖𝑛 − 2𝑉𝑐2

𝐶𝑒2𝑅𝑤𝑐2

(3.12)

3.5 Simulation Results

The theoretical model is simulated with the parameters of AQUA EWP CDI unit mentioned

in the paper [30], [31].

3.5.1 Simulating the experimental AQUA EWP CDI system using the model

The CDI system model is simulated with the following specifications as mentioned in

Table 3.2.

Table 3.2 Parameters used in AQUA EWP CDI simulation model

Number of CDI cells 2

Feed water concentration 1 g/l (1000 mg/liter (ppm))

Dimension of electrodes (𝑊𝑒 ∗ 𝐿𝑒 ∗ 𝑑𝑐) 0.158 m × 0.174 m × 0.0003 m

Separation between electrodes, 𝑑𝑒 0.002 m

Feed water flow rate 𝑞𝑓 2 liters/min

Specific area of electrodes, 𝑘𝑠 800 𝑚2/𝑔𝑟𝑎𝑚

Mass of activated carbon in a CDI cell, 𝑚𝑐1 = 𝐴𝑐1𝑡𝑐𝜌𝑐 1352.6 grams

61

The other parameters of the system used in the model can be computed using the data in

Table 3.2.

The characteristic thickness of the electric double layer is given by debye length that varies

with concentration in the CDI cell.

Inverse Debye length 𝜅 is given as [29] as given in Equation (2.1),

𝜅1,2 = √2𝐹2𝑥1,2

𝜀𝑟𝜀0𝑅𝑇

(3.13)

where 𝑥1 and 𝑥2 are the effluent concentration varying with respect to time in CDI cell 1

and cell 2.

Debye length is given by 𝜆𝐷1,2=

1

𝜅1,2.

The effective capacitance of AC is given as,

𝐶𝑒1,2

=𝜀0𝐾𝑐𝐴𝑐1,2

𝑡𝑐1,2

(3.14)

Where 𝑡𝑐1 and 𝑡𝑐2

are the thickness of double layer in CDI cell 1 and CDI cell 2.

The simulations are performed in MATLAB. A potential of 1.5 Volts DC is applied to the

CDI cells during purification for 90 seconds followed by zero volts for 30 seconds and -

1.5 volts DC for 30 seconds to totally remove the adsorbed ions during purification step.

The plot of effluent concentration varying with respect to time is obtained using the model

and is shown in Figure 3.5. The salt concentration decreases from 1 grams/liter (g/l) to a

minimum of 0.108 g/l at the end of purification phase of 90 seconds, whereas the

62

concentration reaches a maximum of 3.570 g/l during regeneration phase whereas the

experimental results in [30] for one of the cycles showed a minimum concentration of 90

ppm at the end of purification phase and a maximum TDS concentration of 3.400 g/l during

regeneration phase.

Figure 3.5 CDI operational cycle using model- purification and regeneration – showing effluent concentration

with time

The capacitor voltage profile for CDI cell 2 is shown in Figure 3.6. The voltage increases

during charging of the CDI cell upon application of a positive voltage and decreases during

the regeneration step.

63

Figure 3.6 Capacitor voltage profile with time

The charge captured by the CDI cell 2 with respect to time is shown in Figure 3.7. During

charging, as the ions are adsorbed onto the electrodes we see that the charge increases with

time while during regeneration or discharging of the CDI cell, the charge decreases to a

negative value.

Figure 3.7 Captured charge in coulombs in the CDI cell 2

64

The current profile in CDI cell 2 as a function of time is shown in Figure 3.8. It decreases

during charging of a CDI cell. As the concentration reaches the lowest value, the current

decreases to zero.

Figure 3.8 Current in CDI cell 2 varying as a function of time (in amperes)

3.5.2 Investigating removal efficiencies and other operational

parameters

TDS removal efficiency is defined as the ratio of the difference of feedwater concentration,

𝑥𝑖𝑛 and the freshwater concentration, 𝑥𝑝 (or the lowest effluent concentration) to the

feedwater concentration.

TDS removal efficiency = 𝑥𝑖𝑛−𝑥𝑝

𝑥𝑖𝑛× 100 %

Investigation I. TDS removal efficiencies were calculated using the model in Section 3.2

for different concentrations of feed water with a flow rate of 2 liters/minute and temperature

of 24 ℃ and the results are compared with the experimental results of the paper [31]. The

65

TDS removal efficiency is calculated for different concentrations ranging from 0.5 – 2.5

g/l and the results are compared with experimental values as shown in Table 3.3.

Table 3.3 Comparison of TDS removal efficiency between model and experiment [31] for different TDS

concentrations.

Concentration

g/l

TDS removal efficiency

(Model) (%)

TDS removal efficiency [31]

(Experiment) (%)

0.5 94.02 95.26 ± 1.02

1 89.10 87.67 ± 0.55

1.5 83.85 79.54 ± 1.26

2 79 71.43 ± 1.74

2.5 74.35 62.39 ± 0.79

As the feed concentration increases, the TDS of the water coming out during purification

phase increases. Also, the rate of salt removal increases and reaches a constant value as the

electrodes reach their saturation capacity, i.e., no more ions can be adsorbed on to the

electrodes. The thickness of double layer decreases with increase in concentration. In Table

3.3, we see that the electrosorption removal efficiency using the model is very well

described for concentration ranging from 0.5-2 g/l. However, as the concentration

increases, the model predicts decreasing trend in efficiency but does not predict the results

accurately. The inconsistency could be attributed to some ions residing in the electrode

pores during regeneration and not being flushed out so that in the next purification phase

the salt removal is less and the electrodes reach saturation much faster. The electrode

capacity obtained using the model is just an approximation. The actual response of porous

electrodes is nonlinear and needs to be further investigated and incorporated in the model.

66

Investigation II. TDS removal efficiencies were calculated using the model in Section 3.2

for different flow rates of feed water with a fixed TDS concentration of 1 g/l and

temperature of 24 ℃ and the results are compared with the experimental results of the paper

[31]. The comparison of results obtained using theoretical model and experiment is

mentioned in Table 3.5.

Table 3.4 Comparison of effluent TDS concentration of the purified stream between model and experiment [31]

for different flow rates.

Flow rates TDS removal efficiency

(Model) (%)

TDS removal efficiency [31]

(Experiment) (%)

1 92.14 95.08 ± 0.51

1.5 90.61 91.11 ± 0.59

2 89.10 87.68 ± 0.55

2.5 87.55 84.28 ± 0.51

3 85.98 81.41 ± 0.52

3.5 84.40 78.20 ± 0.63

TDS removal efficiency is computed using the model for feed water concentration of 1 g/l

but with different flow rates ranging from 1-3.5 liters/ min at a temperature of 24 ℃. As

the flow rate increases, the water stays in the cell for smaller duration and as a result less

time is available for ions to move to the activated carbon electrodes. Thus, we see there is

a declining trend in TDS removal efficiency with the flow rate which is very well described

using the model.

67

3.6 Investigation of CDI batch mode of operation using model

++ +

- --

u1

-u1

q, xin q, xp

E-1

E-2

E-3

Figure 3.9 CDI Batch mode of operation

In the batch mode of operation, the freshwater obtained at the outlet is recycled back to the

inlet tank [29], [89]. The water from the tank now flows back again into the CDI cell. The

batch mode of operation allows further purification of the produced water. As the

electrodes get saturated with ions, the voltage is removed so that the ions are released into

the solution and the water is discarded away.

3.6.1 CDI model – batch mode

Let 𝑉𝑖𝑛 be the volume of water in the inlet tank, 𝑥𝑖𝑛 be the concentration of the inlet tank.

The flow rate in and out of the CDI cell is same. Let q be the flow rate of water to and out

of the CDI cell and 𝑉𝑡𝑎𝑛𝑘 be the volume of the inlet tank.

68

The salt balance equation of the inlet tank is written as,

𝑉𝑡𝑎𝑛𝑘

𝑑𝑥𝑖𝑛

𝑑𝑡= 𝑞𝑥𝑝 − 𝑞𝑥𝑖𝑛

(3.15)

The state equations describing effluent salt concentration and voltage across capacitor for

CDI batch mode of operation can be written as,

𝑑𝑥𝑝

𝑑𝑡=

𝑞

𝑉(𝑥𝑖𝑛 − 𝑥𝑝) −

𝐾

𝑉

(2𝑉𝑖𝑛 − 2𝑉𝑐)

𝑅𝑤𝑐

(3.16)

𝑑𝑉𝑐

𝑑𝑡=

2𝑉𝑖𝑛 − 2𝑉𝑐

𝐶𝑒𝑅𝑤𝑐

(3.17)

Equation (3.15), (3.16) and (3.17) form the state equations for batch mode CDI operation.

3.6.2 Results – CDI batch mode of operation

The simulation for batch mode operation of AQUA EWP CDI unit is performed. The inlet

water tank is filled with 10 liters of water of concentration 1 g/l. The water flows through

the CDI cell at a flow rate of 2 liters/min. The voltage is applied to the CDI cell until a

constant effluent concentration is reached in the inlet tank. The water is recirculated back

to the tank until a constant effluent salt concentration is reached. As seen in Figure 3.10, it

is observed that the water concentration in the tank decreases from 1 g/l and reaches a

steady concentration of 0.323 g/l at the end of 1000 seconds. This implies it takes about

1000 seconds for the electrodes to reach saturation.

69

Figure 3.10 Effluent concentration at the output of CDI cell and in the inlet tank

From Figure 3.11, it is observed that as the flow rate of water increases, the steady-state

effluent concentration is less compared to the case of low flow rate. This implies that flow

rate is directly proportional to CDI desalination capability.

70

Figure 3.11 Effluent concentration in the inlet tank at different flow rates q

Moreover, it is also observed from Figure 3.11 that the time taken to reach the lowest

concentration in the steady state is small if the flow rate is high. At a flow rate of 4

liters/minute, the steady state concentration of 0.307 g/l is reached after 500 seconds

whereas, at a flow rate of 6 liters/min, a steady state concentration of 0.297 g/l is reached

after 450 seconds. This behavior of batch mode operation to flow rate is attributed to the

mixing activity in the inlet tank as the water is recycled suggesting that lower flow rate

would reduce the flow of the water back into the inlet tank which would stop continuous

removal of ions by CDI cell. As the flow rate increases, a point is reached where the lowest

effluent concentration in the tank no longer decreases. Thus, an optimized flow must be

used for efficient batch mode CDI operation [89].

71

3.7 Conclusion

A mathematical model of capacitive de-ionization system with activated carbon electrodes

is described for water desalination. The model is extended to two CDI cells to model

AQUA EWP CDI unit. The model is used to reproduce the operational cycle of AQUA

EWP CDI cell that shows the effluent concentration varying with time. The model results

are in good agreement with the reported results of AQUA EWP CDI unit. The model also

predicts the amount of charge captured by the CDI cell and the current in the CDI cell, both

as a function of time. The influence of parameters such as flow rate and influent TDS

concentration on the TDS removal efficiency of a CDI system is very well described by

the model. The model predicts decreasing trend in TDS removal efficiency with increase

in feed concentration and flow rate. However, at higher concentrations of the feed, a

deviation from experimental values is observed. The electrode capacity obtained using the

model is just an approximation. However, the actual electrode capacity depends on various

parameters and the relation describing it is highly nonlinear that needs to be further

investigated. The influence of other parameters on the system performance can be studied

by minor additions to the model.

The batch mode of operation of CDI is investigated using the same model and it has been

found that the flow rate has a positive effect on CDI performance. As the flow rate

increases, the lowest effluent salt concentration decreases for the same influent

concentration. This is attributed to the mixing activity in the recycling tank that decreases

the concentration in the tank. Higher flow rate causes more water to circulate back in the

tank and thus the concentration decreases fast. Moreover, the time required to reach steady

72

state decreases with increase in flow rate. As the flow rate is increased, a point is reached

when the lowest effluent salt concentration remains same no matter how much the flow

rate is increased. Therefore, an optimized flow is required for efficient batch mode CDI

operation.

This model can serve as a benchmark to study the performance of CDI system under

different configurations, different operating conditions, understanding the effect of

different parameters on system performance and also optimization of the system

parameters to achieve maximum efficiency.

73

4 CHAPTER 4

SWITCHED WAVE CAPACITIVE DE-IONIZATION

4.1 Introduction

A typical CDI process as described in Chapter 3 consists of 2 steps viz., ion adsorption

wherein freshwater is obtained and ion desorption wherein a highly concentrated brine is

obtained. This chapter proposes a novel approach of continuous mode of operation of

Capacitive De-ionization called Switched wave capacitive de-ionization (SWCDI) wherein

freshwater and brine are collected at the same time. The mathematical model developed in

Chapter 3 is extended to describe SWCDI operation using the concept of switched electric

wave. Three CDI cells in series are considered wherein a moving electric field in the first

cell separates the salt ions by adsorption of cations to the cathode and anions to the anode

further forcing the cations and anions to move to a second CDI cell and finally to the third

one. The flow out of the first cell has a substantially less total dissolved salt (TDS) than the

inlet flow to the cell. The accumulated charged water in the regions of third CDI cell is

recombined and discharged as brine. The brine flow rate is adjusted to achieve an

economical water recovery ratio. The parameters involved in the system are optimized

using a genetic algorithm to achieve desired freshwater concentration and high water

recovery ratio.

This chapter is organized as follows. Section 4.2 describes the instrumentation involved in

the switched wave capacitive de-ionization (SWCDI) for continuous operation of CDI

74

along with overall mass balance and salt balance equations of the system. The description

of Switched wave Capacitive De-ionization system is presented in Section 4.3. Next, a

mathematical model of switched wave capacitive de-ionization (SWCDI) system is derived

using an electrical equivalent circuit of CDI in Section 4.4. Simulation results for SWCDI

system is presented in Section 4.5. In Section 4.6, reflux mode of operation of SWCDI

system is investigated and its results are discussed.

4.2 Description of SWCDI system

The block diagram of proposed switched wave capacitive de-ionization system is as shown

in Figure 4.1.

Brine Tank

Inlet Tank

qb

qpqf

SWCDI Cell

Saline water

q

Figure 4.1 Block diagram of switched wave capacitive de-ionization system

75

The salty water is first collected in an inlet tank, where the inlet flow 𝑞 is controlled by a

standard level control instrumentation. A pump takes the salty water from the inlet tank at

a flow rate of 𝑞𝑓 to the assembly of CDI cells. The CDI cells consist of a main conduit,

preferably of a rectangular shape, made from an electrically insulated material as plastic or

glass. The feed raw water 𝑞𝑓 is supplied through the pump to the Switched Wave

Capacitive De-Ionization (SWCDI) Cell. Across two sides of the main conduit, for

example the upper and lower sides, the electrodes are connected to an electronic voltage

waveform generator. The inlet water 𝑞𝑓 passes into the CDI cells assembly wherein the salt

ions are separated under the applied electric field and attracted towards respective

electrodes. The travelling wave allows the ions to move from one electrode to another and

finally the ions are transported to the final CDI cell, where the cation and anion ions reunite

again in the outlet chamber, where the water becomes neutralized, with high concentration

and leaves the conduit through the brine output pipe. The brine discharge rate 𝑞𝑏 can be

adjusted to produce the desired water recovery ratio. The low salt water flow flows through

a separate pipe from the output of first CDI cell with a flow rate 𝑞𝑝 .

At steady state, the overall mass balance equation of the entire system is given by

𝑞𝑓 = 𝑞𝑏 + 𝑞𝑝 (4.1)

Where

𝑞𝑓 is the salty water inlet volume feed rate (𝑚3/sec)

𝑞𝑝 is the permeate (distillate) volume flow rate

𝑞𝑏 is the brine volume flow rate.

76

The salt balance equation is given as,

𝑥𝑓𝑞𝑓 = 𝑥𝑝𝑞𝑝 + 𝑥𝑏𝑞𝑏 (4.2)

Where

𝑥𝑓 is the feed water concentration in 𝑔/𝑙𝑖𝑡𝑒𝑟

𝑥𝑝 is the permeate salt concentration in 𝑔/𝑙𝑖𝑡𝑒𝑟.

𝑥𝑏 is the brine salt concentration in 𝑔/𝑙𝑖𝑡𝑒𝑟.

The recovery ratio is given by 𝑅𝑅 =𝑞𝑝

𝑞𝑓.

which implies that salt balance equation can be written as,

𝑥𝑓 = 𝑥𝑏(1 − 𝑅𝑅) + 𝑥𝑝𝑅𝑅 (4.3)

77

4.3 Switched Wave Capacitive De-ionization (SWCDI) cell

+ + +

- --

qo

qb

qin

-u1

+u3

+u1

-u3

sw1 sw2 sw3sw4

sw1 sw2 sw3sw4

Note: The lower part is simply a mirror

image of the upper part

Figure 4.2 Assembly of CDI cells arrangement in a SWCDI cell

The SWCDI cell is a conduit of rectangular cross section of width 𝑊𝑒, length 𝐿𝑒 and height

𝑑𝑒. As shown in Figure 4.2, a SWCDI conduit carries a series of electrodes on two opposite

sides. Each electrode of a CDI cell consists of a conducting charge collector and an

activated carbon (AC) layer sheet of thickened 𝑡𝑐. The activated carbon layer merged and

fully in contact with the salty water. The top metal electrode in the first CDI cell is

connected to electrical signal 𝑢1+, while the lower metal electrode is connected to electrical

signal 𝑢1−. The switching action determines the charging and discharging of the

neighboring cells.

78

The capacitance associated with each electrode is given by 𝐶𝑒 =𝑘𝑐𝜀0(𝐿𝑒𝑊𝑒)

𝑡𝑐𝑠, where 𝑘𝑐

depends on the effective area ratio of the activated carbon and its wet permittivity, tcs is the

thickness of the Stern layer which is of the order of 1 nanometer.

𝑘𝑐 = 𝑘𝑠𝜌𝑐𝑡𝑐; where 𝑘𝑠 is the specific surface of the AC. 𝑘𝑠 is typically 1000-2500 𝑚2/𝑔.

The Table 4.1 below demonstrates the working principle of SWCDI cell.

Table 4.1 Switching time and operation in SWCDI cell

Timeslot

(seconds)

Sw1 Sw2 Sw3 Sw4

𝑇𝑠1 On Off On Off

𝑇𝑠2 Off On Off On

𝑇𝑠3 On Off On Off

𝑇𝑠4 Off On Off On

The principle of operation of SWCDI cell is as follows as shown in Table 4.1.

1. Initially, the CDI cell 1 is charged, that results in ion adsorption onto the electrodes.

2. The charge on CDI cell 1 is transferred to CDI cell 2, that results in ion desorption

in CDI cell 1. The desorbed ions are adsorbed onto the CDI cell 2.

3. The charge on CDI cell 2 is transferred to CDI cell 3, that results in ion desorption

from CDI cell 2. The desorbed ions are adsorbed onto CDI cell 3.

4. CDI cell 3 is discharged by a negative voltage 𝑢3 that results in flushing of all the

ions with a brine flow.

79

The steps are performed in parallel, steps 1,3 in one-time slot (𝑇𝑠1) and steps 2,4 in the next

time slot. (𝑇𝑠2). These steps are repeated until a steady state of the system is reached.

The equivalent circuit of the six electrodes during 𝑇𝑠1 and 𝑇𝑠2

is shown in Figure 4.3.

Rs1

Rw1

Rs2

Rw3

C3C2C1

C3C2C1

Rs1Rs2

-U1 -U3

Sw1 Sw2 Sw3 Sw4

Sw1 Sw2 Sw3 Sw4

U1 U3

Figure 4.3 Equivalent electrical circuit of 3 CDI cells in series

𝑅𝑤 is the electrical resistance of the salty water between the two opposite electrodes, while

𝑅𝑠 is the electrical resistance between two adjacent electrodes.

The resistance of water 𝑅𝑤 and resistivity is given by the equation (3.3) and (3.4).

The series resistance 𝑅𝑠 governs the movement of ions from one CDI cell electrode to the

adjacent electrode. The series resistance 𝑅𝑠 is given by the approximation,

𝑅𝑠 =

𝜌𝑒𝑑𝑡

𝑊𝑒𝑡𝑐𝑘𝑝

(4.4)

Where 𝑘𝑝 is the shaping factor,

80

𝜌𝑒 is water resistivity,

𝑑𝑡 is the distance between adjacent electrode,

𝑊𝑒 is width of the electrode, and

𝑡𝑐 is the thickness of Activated carbon layer.

During normal operations, 𝑥𝑓 and RR are constants and 𝑞𝑓 is adjusted to achieve the

desired freshwater salt concentration (say 500 ppm).

Using equation (4.1), 𝑞𝑝 = 𝑞𝑓 ∗ 𝑅𝑅 , and 𝑞𝑏 = 𝑞𝑓 ∗ (1 − 𝑅𝑅).

4.4 Modelling of SWCDI cell

Let 𝑉𝑐𝑖 be the voltage across the ith capacitor, and 𝐼𝑠𝑖

be the current throw the 𝑅𝑠𝑖 resistance

and 𝐼𝑤𝑖 be the current going throw 𝑅𝑤𝑖 resistance (i=1,2,3).

The values of 𝑉𝑐𝑖, 𝐼𝑤𝑖

and 𝐼𝑠𝑖 are calculated for two different timeslots 𝑇𝑠1

and 𝑇𝑠2 as follows.

81

During 𝑻𝒔𝟏:

Rs1

Rw1

Rs2

Rw3

C3C2C1

C3C2C1

Rs1 Rs2

-U1

U1

Is1 Is2

Is1 Is2Iw1 Iw3

+Vc1

-

+Vc2

-

+Vc3

-

_Vc2

+

_Vc1

+

_Vc3

+

Figure 4.4 Equivalent electrical circuit during 𝑻𝒔𝟏

From the equivalent electrical circuit of CDI cell in Figure 4.3, we write the following

equations,

𝐼𝑤1

=2

𝑅𝑤1

(𝑢1 − 𝑉𝑐1)

(4.5)

Applying KVL,

𝐼𝑠2

=1

𝑅𝑠2

(𝑉𝑐2− 𝑉𝑐3

) (4.6)

𝐼𝑐3= − 𝐼𝑐2 (4.7)

82

Applying KVL to inner loop,

−𝐼𝑤3𝑅𝑤3

+ 2𝐼𝑠1𝑅𝑠1

+ 2𝐼𝑠2𝑅𝑠2

+ 𝐼𝑤1𝑅𝑤1

= 0 (4.8)

𝐼𝑐1

= 𝐶1

𝑑𝑉𝑐1

𝑑𝑡= 𝐼𝑤1

− 𝐼𝑠1

(4.9)

𝐼𝑐2

= 𝐶2

𝑑𝑉𝑐2

𝑑𝑡= 𝐼𝑠1

− 𝐼𝑠2

(4.10)

𝐼𝑐3

= 𝐶3

𝑑𝑉𝑐3

𝑑𝑡= 𝐼𝑠2

+ 𝐼𝑤3

(4.11)

From Equations (4.7), (4.10) and (4.11),

𝐼𝑤3= −𝐼𝑠1

(4.12)

From Equations (4.8), (4.12),

𝐼𝑤3 =

1

𝑅𝑤3+ 2𝑅𝑠1

(2𝑅𝑠2𝐼𝑠2

+ 𝐼𝑤1𝑅𝑤1

) (4.13)

Equations (4.5), (4.6), (4.12), (4.13) provides the values of 𝐼𝑤1,𝐼𝑠2, 𝐼𝑠1

𝑎𝑛𝑑 𝐼𝑤3 while

Equations (4.9), (4.10) and (4.11) provide state equations for capacitor voltages.

83

During 𝑻𝒔𝟐:

Rs1

Rw1

Rs2

Rw3

C3C2C1

C3C2C1

Rs1 Rs2

-U3

U3

Iw1 Iw3

Is1 Is2

Is1 Is2

Figure 4.5 Equivalent electrical circuit during 𝑻𝒔𝟐

𝐼𝑠1

=1

𝑅𝑠1

(𝑉𝑐1− 𝑉𝑐2

) (4.14)

𝐼𝑐1= − 𝐼𝑐2 (4.15)

𝐼𝑤3

=1

2𝑅𝑤3

(𝑢3 − 𝑉𝑐3)

(4.16)

From Equations (4.9), (4.10), (4.11) and (4.15),

𝐼𝑠2= 𝐼𝑤1

(4.17)

From Equation (4.8),

𝐼𝑤1

=1

𝑅𝑤1+ 2𝑅𝑠2

(−2𝑅𝑠1𝐼𝑠1

+ 𝐼𝑤3𝑅𝑤3

) (4.18)

84

Equations (4.14), (4.16), (4.17) and (4.18) give 𝐼𝑠1, 𝐼𝑤3

, 𝐼𝑠2, 𝐼𝑤1

respectively while

Equations (4.11), (4.12) and (4.13) provide state equations for capacitor voltages.

The concentration, 𝑥1 and 𝑥3 in the first and last chamber of the SWCDI cell is calculated

using the differential equation,

𝑑𝑥1

𝑑𝑡= 𝑠𝑛1

𝐾

𝑉1𝐼𝑤1

+𝑞𝑓𝑥𝑓

𝑉1−

𝑞𝑝𝑥1

𝑉1

(4.19)

𝑑𝑥3

𝑑𝑡= 𝑠𝑛3

𝐾

𝑉3𝐼𝑤3

+𝑞𝑏

𝑉3(𝑥𝑓 − 𝑥3)

(4.20)

Where = 𝑀𝑤𝑠/ 𝐹 ; and 𝑀𝑤𝑠 is the molecular weight of salt, and F is Faraday's constant.

𝑠𝑛𝑖= ±1 depending on 𝑉𝑐𝑖

𝑎𝑛𝑑𝑑𝑉𝑐𝑖

𝑑𝑡 as illustrated in the table.

Table 4.2 Sign of 𝒔𝒏𝒊for 𝑽𝒊 and 𝑽𝒊

𝑉𝑖 > 0, 𝑉�� > 0 𝑠𝑛𝑖= −1

𝑉𝑖 > 0, 𝑉�� > 0 𝑠𝑛𝑖= +1

𝑉𝑖 < 0, 𝑉�� < 0 𝑠𝑛𝑖= −1

𝑉𝑖 < 0, 𝑉�� > 0 𝑠𝑛𝑖= +1

The above equations describe the relationship between currents and voltages for each stage,

as well as the concentration of the three regions.

85

𝑥𝑝 = 𝑥1; 𝑥𝑏 = 𝑥3 (4.21)

where 𝑥𝑝 is the freshwater concentration coming out of the SWCDI cell and 𝑥𝑏 denotes

the brine concentration coming out of SWCDI cell

4.5 Optimization algorithm to achieve desired response

The flow rate has significant effect on CDI process as mentioned in Chapter 3. Lower the

flow rate, higher the salt removal efficiency as the ions get sufficient time to move to

respective electrodes inside the CDI cell.

The distance between the electrodes is another significant parameter affecting CDI

performance. As the voltage is applied between the electrodes, an electrostatic field is

generated that causes the ions to move towards oppositely charged electrodes. This

electrostatic field decreases as the spacing is increased corresponding to low salt removal.

The spacing between the electrodes must be small but it must be made sure that the

activated carbon on the top and bottom electrodes are not in contact with each other which

would otherwise result in short-circuiting of the CDI cell.

The separation between adjacent electrodes is also critical for movement of ions from one

CDI cell electrode to the next. Smaller the distance, higher is the rate of transfer of ions

under the effect of switched configuration. All these parameters are therefore optimized

for efficient SWCDI operation.

86

The desired objective is to obtain freshwater with low concentration of salts (500 ppm or

0.5 grams/liter) and to increase the water recovery ratio (RR). The initial feed concentration

is 1 g/l. To achieve the required objectives, a genetic optimization algorithm is run to

minimize the following objective function

𝐽 = (𝑥1 − 𝑥𝑑)2 + (

50

𝑞𝑓) + (

1

𝑅𝑅)

(4.22)

Where 𝑥1 is the freshwater concentration varying with time according to Equation (4.14),

𝑥𝑑 is the desired freshwater concentration, RR is the water recovery ratio.

An optimization algorithm is run in MATLAB to obtain the following optimized set of

parameters for the SWCDI cell: Switching times 𝑇𝑠1 𝑎𝑛𝑑 𝑇𝑠2

, negative voltage 𝑢3, feed

flow rate 𝑞𝑓, spacing between electrodes of a CDI cell 𝑑𝑒, spacing between adjacent

electrodes in SWCDI cell 𝑑𝑡 and water recovery ratio 𝑅𝑅. The other parameters of the

SWCDI system are kept constant during the optimization as shown in Table 4.2.

4.6 Simulation Results

The feed solution considered is sodium chloride 𝑁𝑎𝐶𝑙 of 1 g/l concentration. The three

zone SWCDI system is simulated with the following parameters in Table 4.3.

87

Table 4.3 Parameters of SWCDI cell

Feed water concentration 1 g/l

Length of each electrode, 𝐿𝑒 0.3 m

Width of each electrode, 𝑊𝑒 0.3 m

Specific surface area of AC, 𝑘𝑠 2000 𝑚2/𝑔𝑟𝑎𝑚

Thickness of the Activated carbon film 0.5 mm

Applied voltage, 𝑢1 1.2 volts

Molecular weight NaCl, 𝑀𝑤𝑠 58.44 grams

During optimization, the parameters in Table 4.3 are kept constant and the parameters in

Table 4.4 are varied by an optimization routine to minimize the cost function J in equation

(4.17). The optimized set of parameters of SWCDI system is as shown in Table 4.4.

Table 4.4 Optimized parameters of SWCDI system

Separation between adjacent electrodes, 𝑑𝑡 0.001 m or 1 mm

Separation between electrodes, 𝑑𝑒 0.008 m or 8 mm

Switching times, 𝑇𝑠1 and 𝑇𝑠2

344 seconds

Feed water flow rate, 𝑞𝑓 4.5 ml/min

Water recovery ratio, 𝑅𝑅 0.73

Voltage to CDI cell 3, 𝑢3 -0.3 volts

88

The saline water concentration change in the first CDI cell and the last CDI cell of SWCDI

system is shown in Figure 4.4. A decrease in salt concentration in the first CDI cell is

observed with time whereas there is an increase in salt concentration in third CDI cell

owing to movement of ions from first to third CDI cell. The ions accumulated in the third

CDI cell is discharged with a flow rate of 𝑞𝑏 as highly concentrated brine whereas the low

concentration water out of the first CDI cell is collected as freshwater.

Figure 4.6 Concentration variation in SWCDI cell

The voltage across capacitor (𝑉𝑐1, 𝑉𝑐2 𝑎𝑛𝑑 𝑉𝑐3

) varies in CDI cell 1,2 and 3 as shown in

Figure 4.5. The voltage across capacitor increases during charging of a CDI cell and

decreases when the cell is discharged.

89

Figure 4.7 Voltage across capacitor in SWCDI cell

4.6.1 Discussion

All the parameters in Table 4.3 are optimized to achieve desired response of the SWCDI

cell. Simulations are performed using the optimized set of parameters. As shown in Figure

4.4, the concentration in the first CDI cell decreases from 1 g/l to 0.5 g/l whereas the brine

concentration in the last CDI cell increases from 1 g/l to 2.2 g/l.

90

4.7 Investigation of reflux mode of operation of SWCDI

The reflux mode of operation of SWCDI cell is investigated in this study wherein a fraction

of low salt water is recycled back to the inlet tank for controlling the amount of salt in

produced freshwater.

4.7.1 Description of reflux SWCDI

Brine TankInlet Tank

qb

qp

qf

qin

qR

qo

SWCDI Cell

Figure 4.8 Diagram of SWCDI system with reflux

The SWCDI cell with reflux flow is shown in Figure 4.6. The description is same as in

Section 4.2 except that the low salt water output 𝑞𝑜 here is split into two flows, a permeate

flow 𝑞𝑝, and reflux flow 𝑞𝑅 . The reflux flow is governed by a pump or a control valve. The

reflex 𝑞𝑅 is recombined with the inlet salty water for controlling the amount of salt in the

produced freshwater 𝑞𝑝.

In order to model the system, additional modelling of the inlet tank is required.

91

4.7.2 Inlet tank model

The mass balance equation for the inlet tank is given by,

𝑑𝑉𝑖𝑛

𝑑𝑡= 𝑞𝑓 + 𝑞𝑅 − 𝑞𝑖𝑛

(4.23)

Where

𝑉𝑖𝑛 is the volume of water in the inlet tank, 𝑞𝑅 is the reflex volume flow rate, and the

change in the streams densities is neglected.

The reflux ratio is defined by the parameter,

𝛼𝑅 =𝑞𝑅

𝑞𝑓 (4.24)

At steady state, the mass balance equation is written as

𝑞𝑓 + 𝑞𝑅 = 𝑞𝑖𝑛 (4.25)

or 𝑞𝑖𝑛 = (1 + 𝛼𝑅)𝑞𝑓 (4.26)

The salt balance equation of the inlet tank is given as,

𝑑(𝑉𝑖𝑛𝑥𝑖𝑛)

𝑑𝑡= 𝑞𝑓𝑥𝑓 + 𝑞𝑅𝑥𝑝 − 𝑞𝑖𝑛𝑥𝑖𝑛

(4.27)

If the inlet tank volume is regulated by a separate loop control, then 𝑉𝑖𝑛 can be considered

constant. Thus, equation can be written as,

𝑉𝑖𝑛

𝑑𝑥𝑖𝑛

𝑑𝑡= 𝑞𝑓𝑥𝑓 + 𝑞𝑅𝑥𝑝 − 𝑞𝑖𝑛𝑥𝑖𝑛

(4.28)

92

At steady state,

𝑞𝑖𝑛𝑥𝑖𝑛 = 𝑞𝑓𝑥𝑓 + 𝑞𝑅𝑥𝑝 (4.29)

Or 𝑥𝑖𝑛 =

𝑥𝑓 + 𝛼𝑅𝑥𝑝

1 + 𝛼

(4.30)

4.7.3 Modelling of reflux SWCDI

The modelling equations are same as in Section from Equation (4.7) – (4.13). The

concentration in the first and last chamber, 𝑥1 and 𝑥3 is calculated using the differential

equation,

𝑑𝑥1

𝑑𝑡= 𝑠𝑛1

𝐾

𝑉1𝐼𝑤1 +

𝑞𝑖𝑛𝑥𝑖𝑛

𝑉1−

𝑞𝑝𝑥1

𝑉1−

𝑞𝑅𝑥1

𝑉1

(4.31)

𝑑𝑥3

𝑑𝑡= 𝑠𝑛3

𝐾

𝑉3𝐼𝑤3 +

𝑞𝑏

𝑉3(𝑥𝑓 − 𝑥3)

(4.32)

𝑥𝑖𝑛 =

𝑥𝑓 + 𝛼𝑥𝑝

1 + 𝛼; 𝑥𝑝 = 𝑥1; 𝑥𝑏 = 𝑥3

(4.33)

4.7.4 Results – Effect of reflux ratio on permeate concentration

We investigate the effect of reflux flow rate (or reflux ratio - Alpha) on the system

performance. The recovery ratio is set to 80 % and feed water flow rate 𝑞𝑓 to 0.15

liters/min. The other parameters used in the simulation are given in Table 4.5.

93

Table 4.5 Parameters used in simulation of reflux SWCDI system

Feed rate 0.15 liters/minute

Feed concentration, 𝑥𝑓 4 grams/liter

Volume of the inlet tank, 𝑉𝑖𝑛 20 liters

Discharging voltage, 𝑢3 -0.8 volts

Switching time, 𝑇𝑠1,2 60 seconds

Total cycle time, 𝑇𝑝 𝑇𝑠1+ 𝑇𝑠2

Length of each electrode, 𝐿𝑒 0.4 m

Width of each electrode, 𝑊𝑒 1 m

94

Figure 4.9 Variation of permeate concentration with time for 𝜶𝑹 = 𝟏 and 𝜶𝑹 = 𝟏𝟎

Using reflux ratio of 𝛼𝑅 = 1 as shown in Figure 4.9, we obtain a minimum permeate

concentration of 1.832 grams/liter and the concentration of solution in the inlet tank is

2.916 grams/liter, but as 𝛼𝑅 is increased to 10, the permeate concentration increases

marginally to 1.857 grams/liter and the concentration in the inlet tank decreases to 2.05

grams/liter.

As shown in Figure 4.10, due to mixing in the inlet tank, the concentration reduces from 4

g/l to around 2.1 g/l during SWCDI operation.

95

Figure 4.10 Concentration in the inlet tank

The results are illustrated in Table 4.6 that compares the permeate concentration, tank

concentration and brine concentration for values of 𝛼𝑅 = 4 and 𝛼𝑅 = 10.

Table 4.6 Permeate, Inlet and brine concentration for different reflux ratio

𝜶𝑹 = 𝟒 𝜶𝑹 = 𝟏𝟎

𝒙𝒑 =0.7733 g/l

𝒙𝒊𝒏 =1.4193 g/l

𝒙𝒃 = 24.8903 g/l

𝒙𝒑 =0.77340

𝒙𝒊𝒏 = 1.0680

𝒙𝒃 = 24.6559

From Table 4.6, it is noticed that as 𝛼 increases, 𝑥𝑝 increases but 𝑥𝑖𝑛decreases. This implies

that as alpha increases, SWCDI desalination performance decreases. This could be

96

attributed to the fact that higher flow rate deteriorates the desalting capability of CDI cell

as less time is available for ions to adsorb on the electrodes.

It is also observed that if the volume of tank 𝑉𝑖𝑛 is increased, the concentration in the inlet

tank, 𝑥𝑖𝑛decreases less rapidly with time. It takes a long time to reach a state where the

concentration variation with respect to time is negligible. Because of large volume of the

tank, the concentration reduction is less rapid due to mixing of feed water and reflux water.

4.8 Conclusion

A switched travelling wave capacitive de-ionization system is described for water

desalination wherein inlet brackish water is subjected to an electrostatic field which causes

the positive and negative ions to move to respective electrodes and the switching action

further causes these ions to move from one electrode to other until the ions reach the last

electrode. The positive and negative ions are combined and discharged as brine whereas a

fraction of low salt water is collected and stored in a tank. There is decrease in

concentration in first CDI cell and increase in concentration in the last CDI cell owing to

movement of ions from first CDI cell to last CDI cell. The parameters of SWCDI cell are

optimized to achieve desired concentration and higher water recovery using an

optimization algorithm. The brine flow rate is also adjusted to obtain high recovery ratio.

The effect of reflux on SWCDI operation is studied and it is observed that reflux has no

critical role in desalination capability as it increases the flow rate to the SWCDI cell. Higher

flow rate causes poor desalination capability resulting in very few ions being removed from

the feedwater.

97

5 CHAPTER 5

CDI EXPERIMENTS

5.1 Introduction

A Capacitive De-ionization cell is built using the following components:

1. Acrylic sheet to support the system

2. Aluminium sheets that serve as current collectors

3. Zorflex Knitted FM-10K Activated Carbon cloth

4. Loctite Water proof silicone to seal off the edges

5. Nuts and screws

6. Power supply to supply a voltage of 1.2 volts to the CDI cell

7. Valves and pipes

8. Water tank

The CDI cell is built using acrylic glass that houses the aluminium sheets and activated

carbon cloth. The activated carbon cloth is glued onto the aluminium sheet. A spacing of 5

mm between the two activated carbon cloth is achieved using side acrylic plates.

A voltage of 1.2 volts dc is applied to the CDI cell as shown in Figure 5.1 that causes the

anions in water to move towards anode and cations to move towards the cathode. The ions

are adsorbed on the activated carbon electrodes during adsorption that results in freshwater

production with less TDS compared to the feed. A salinity sensor is used to measure the

concentration of the solution. As the electrode pores get saturated with ions, the effluent

98

concentration begins to rise and the voltage is removed from the CDI cell so that the ions

get desorbed from the electrodes. This step is known as desorption. A typical CDI process

comprises of two-steps viz., adsorption (when the voltage is applied) and desorption (when

the voltage is removed).

++ +

- --

Lede

qo

dc

DC

qin

Figure 5.1 Diagram illustrating CDI operation

5.2 Zorflex Knitted FM-50K Activated Carbon cloth (ACC)

FM-50K ACC [100] is manufactured by the company Calgon carbon. It has an extremely

large surface area of 2000 𝑚2/𝑔𝑟𝑎𝑚. Surface density for FM50K though is nominally 130

𝑔/𝑚2. Higher specific surface area implies, more ions can be adsorbed on the activated

carbon cloth which means more purification.

Zorflex Knitted Activated carbon cloth FM-50k is resistant to bacteria, fungus and molds,

fibrous, flexible, non-toxic and light weight. Figure 5.2 shows FM-50K ACC attached to

an aluminium sheet that serves as electrode for CDI experiments.

99

Figure 5.2 FM-50K Activated carbon cloth attached to the aluminium sheet

The efficiency and adsorption capacity of knitted activated carbon is less affected by pre-

adsorbed moisture than granular activated carbon. The main applications of this cloth is for

water purification, air purification and as electrodes, wound dressing, protective clothing

as socks and gloves.

5.3 Experimental setup of a CDI cell

The experimental prototype is built using an acrylic sheet of 2.8 mm thickness. As shown

in Figure 5.3, a single CDI cell consists of acrylic of dimension 30 cm x 15 cm x 0.5 cm

(length x width x height).The aluminium plate is glued onto the acrylic sheet on the

opposite sides. Activated Carbon cloth is attached to the aluminium sheets using double

sided tape. The metallic nuts and screws hold the aluminium and activated carbon cloth

and provide means to apply a voltage to the CDI cell. The nuts and screws are so attached

to maintain isotropic conductivity vertically i.e., along the aluminium sheet and activated

carbon cloth such that voltage applied to aluminium sheet charges and discharges the

activated carbon cloth respectively. Two holes are drilled one on the top sheet and other at

the bottom sheet to allow water to enter and exit the CDI cell as shown in Figure 5.2. A

100

spacing of 0.5 cm between the activated carbon electrodes is achieved using side acrylic

sheets.

24 cm 3 cm3 cm

0.5 cm

Saline Water Inlet

Water outlet

Aluminium plate

Activated carbon clothAcrylic glass

Metallic Nuts and screws

Figure 5.3 Illustration of experimental setup of a CDI cell

5.3.1 Description of the experimental setup

Figure 5.4 CDI Experimental setup

101

A prototype of CDI is built as shown in Figure 5.4. A DC regulated power supply is used

to apply a fixed voltage of 1.2 volts during the adsorption step.

CDI cell

Brine Tank

qbqp

qf

Inlet

Tank

Fresh water Tank

Salty water

Figure 5.5 Instrumentation of CDI experimental prototype

The piping and instrumentation diagram for a single CDI cell is as shown in Figure 5.5.

The saline water is stored in the inlet tank. The flow of saline water into the CDI cell is

controlled using a valve. As the water flows into the CDI cell, a voltage of 1.2 volts dc is

applied that polarizes the Activated carbon cloth which causes the ions in the water to move

towards oppositely polarized electrodes (anions towards anode and cations towards

cathode). During purification, the valve at the outlet is closed so as not to allow any water

to flow outside the cell. As sufficient ions are removed from the water, the outlet

purification valve is opened to exit freshwater out of the CDI cell.

The ions that are still present in the electrode pores are removed by filling the CDI cell

with water again and short-circuiting the two electrodes that cause the ions to desorb from

102

the electrodes and flow back into the water. This results in highly concentrated stream

(brine) that is discharged using brine discharge valve into the brine tank.

5.3.2 Results and Observation

The saline water with an initial concentration of 5500 ppm (5.5 g/l) is fed to the CDI cell.

The water is purified during adsorption step and reaches a concentration of 4500 ppm (4.5

g/l) upon application of 1.2 volts to the system. During desorption, the CDI cell is again

filled with water and the voltage is removed that causes ions to flush back in water. This

water is discharged as brine. We noticed a brine concentration of 5500 ppm (5.5 g/l) for

same concentration of feed water. Water recovery can be adjusted by varying the flow rates

accordingly. Low flow rates enhance the performance of CDI.

The water discharged as freshwater can be further purified by feeding it back to the inlet

tank. This would imply the batch mode of operation of CDI.

During adsorption, the current to the CDI cell is measured to be 32 mA and decreases to

25 mA in 30 mins. The power consumption is calculated only during ion adsorption step

in CDI as no voltage is applied during ion desorption step. The power consumption is

calculated to be 0.0384 watts.

Precautions:

Care must be taken while handling Activated Carbon cloth as it removes oxygen from the

air. It must be made sure that the activated carbon cloth on the opposite sides must not be

in contact with each other which would result in shorting of a CDI cell.

103

5.4 Experimental setup – SWCDI

dt

Lt

Lede

dt

Salt Water Inlet

Brine outlet

Activated Carbon clothAluminium plate Acrylic glass

Insulator Plastic screwFresh water outlet

Figure 5.6 SWCDI experimental setup

The switched wave capacitive de-ionization requires 3 CDI cells in series as explained in

Chapter 4. A diagram illustrating experimental SWCDI system with all its parts is shown

in Figure 5.6. The experimental prototype of SWCDI consists of acrylic sheets of

dimension 1.2 m x 0.5 m x 0.003 m (𝐿𝑡 ∗ 𝑊𝑡 ∗ 𝑑𝑡). The aluminium and activated carbon

cloth is cut in the dimension 0.3 m x 0.3 m (𝐿𝑒 ∗ 𝑊𝑒). The dimensions of the SWCDI system

is mentioned in Table 5.1.

Table 5.1 Dimension of SWCDI system

Length of acrylic sheet, 𝐿𝑡 1.2 m

Width of acrylic sheet, 𝑊𝑡 0.5 m

Thickness of acrylic sheet, 𝑑𝑡 0.003 m

Length of ACC, 𝐿𝑒 0.3 m

104

Width of ACC, 𝑊𝑒 0.3 m

Thickness of ACC, 𝑡𝑒 0.5 mm

Figure 5.7 depicts aluminium sheets attached to acrylic sheet that serve as the current

collector in SWCDI system.

Figure 5.7 Aluminium sheets attached to the Acrylic sheet

Activated carbon cloth (FM-50k) is attached to aluminium using 3M-9713 double sided

electrical conductive tape [101] as shown in Figure 5.8. The spacing of 𝑑𝑡 = 0.002 m is

maintained between adjacent activated carbon cloths as shown in Figure 5.8. The spacing

between the top and bottom electrodes is maintained to be 6 mm. An insulator material is

fitted in between the activated carbon electrodes in the second CDI cell using plastic nuts

and screws that cancels the effect of electrostatic field in the second CDI cell as shown in

Figure 5.9. As the water flows through the inlet, the ions are adsorbed in the first CDI cell.

The ions move through water from first CDI cell to the last CDI cell under switching

operation of SWCDI system as explained in Section 4. As a result, the freshwater is

collected from first CDI cell and brine is collected from last CDI cell. The experimental

prototype is as illustrated in Figure 5.7-5.9.

105

Figure 5.8 Activated carbon attached to the aluminium sheets using 3M-9713 Electrical conducting tape

Figure 5.9 SWCDI system with connections to switches

5.4.1 Observations and Recommendation

The activated carbon particles often end up in the 2 mm space between the adjacent

electrodes that forms an electrical contact between the adjacent electrodes. This can be

avoided by using an insulator material between adjacent electrodes. The ions moving from

first CDI cell to next is dependent on the series resistance 𝑅𝑠1,2. Lower the resistance, more

106

will be the ion transport from one CDI cell to the next. The resistance can be reduced by

employing different geometries of activated carbon electrodes.

Another concern is the time constant of the system which is given by 𝜏 = 𝑅𝐶. The

capacitance is very high because of the specific surface area of activated carbon electrodes.

To have a small time constant for the system, the resistance must be small which would

imply more transport of ions from one CDI cell to the next in small duration.

The voltage to the first CDI cell is ‘V’ volts. Upon switching action, the next adjacent CDI

cell is charged with a voltage ‘V/2’ volts. A more general switching scheme can be used.

A possible recommendation would be to use an overlapping waveform i.e., before one CDI

cell gets discharged, the neighboring cell is charged so as to allow ions in the water top

move to the neighboring cell with the water flow rather than being flushed back completely

into the water.

All these factors are essential for proper functioning of SWCDI system and requires further

investigation.

107

6 CHAPTER 6

CONCLUSIONS AND FUTURE WORK

6.1 Conclusions

This thesis presents modelling and control of capacitive de-ionization system. A

mathematical model is presented using an electrical equivalent circuit of CDI along with a

dynamic equation that describes how the effluent concentration varies with respect to time

in a CDI cell. The modelling is done considering various parameters associated with a CDI

process viz., the specific surface area of activated carbon electrodes, dimensions of the

electrodes, spacing between the electrodes, solution resistance between the electrodes,

initial feed concentration, flow rate, temperature. The effectiveness of the model to

describe CDI process is evaluated by comparing its results with reported experimental

results of AQUA EWP CDI unit. Simulations were carried out to compute TDS removal

efficiency for different initial feed concentrations of water and flow rates. The model

results showed good agreement with the experimental results. The model predicts

decreasing salt removal efficiency with increase in feedwater salt concentration as

electrodes reach their saturation capacity i.e., they are fully filled with ions. At higher

concentrations, the model does not predict the results accurately, that is attributed to

nonlinear characteristics of porous electrodes that must be incorporated in the model. It is

also observed that salt removal efficiency decreases with increase in flow rate, that is

attributed to low residence time for ions at high flow rates. The batch mode of AQUA EWP

108

CDI unit is investigated using the mathematical model developed for a CDI cell. It is

observed that the effect of flow rate on the batch mode CDI operation is different from the

single flow mode. Here, the salt removal efficiency increases with increase in flow rate due

to the mixing that occurs in the inlet tank and as more water is circulated back to the tank.

The thesis also proposes a novel operational mode that allows continuous mode operation

of CDI cell using the concept of switched electric wave, wherein we obtain continuous

supply of freshwater and brine with time. The model for a single CDI cell is extended to

model Switched Wave Capacitive De-ionization Cell (SWCDI). The simulation results

demonstrate decrease in concentration in the first CDI cell and increase in concentration in

the last CDI cell of SWCDI with respect to time. The system parameters are optimized

using an objective function to achieve desired response i.e., desired freshwater

concentration of 500 ppm and high water recovery ratio.

A reflux mode of SWCDI is also investigated, wherein it is observed that reflux has no

critical effect on CDI desalination performance, as with the increase in reflux rate, the flow

to the SWCDI cell is also increased due to balance equations. This would result in low

desalination performance of SWCDI cell, as flow rate weakens the desalination capability

of the CDI cell.

Lastly, a prototype of a CDI cell is developed wherein it is observed that the salt

concentration decreases under the effect of applied potential to the electrodes. A prototype

of SWCDI system is also developed to demonstrate its ability to produce freshwater and

brine simultaneously and the critical factors for efficient functioning of the SWCDI system

are studied.

109

6.2 Future Work

The following are the recommendations for further research in this work:

1. Incorporate the nonlinear model describing electrical double layer in the developed

mathematical model.

2. Use of overlapping voltage waveforms in SWCDI system.

3. Investigate different schemes of CDI using the developed model and study

dependence of parameters on the system performance.

4. Integrate the model with energy converter that recovers the energy released during

desorption in a CDI cell and uses it to charge a neighboring CDI cell or a battery.

5. Solar powered CDI devices can be manufactured that can be deployed to a location

on demand that would save power consumption of CDI cell.

6. Extend the model to describe Solar based CDI.

7. Extend the model to describe the hybrid mode of CDI with other desalination

systems.

110

References

[1] R. Zhao, Theory and operation of capacitive deionization systems. 2013.

[2] “Water Crisis | Water Ambassadors Canada.” [Online]. Available:

http://www.waterambassadorscanada.org/water_crisis.html. [Accessed: 04-Apr-

2015].

[3] R. Zhao, P. M. P. Biesheuvel, A. van der Wal, and A. Van der Wal, “Energy

consumption and constant current operation in membrane capacitive deionization,”

Energy & Environmental Science, vol. 5, no. 11, p. 9520, 2012.

[4] P. Gleick, “Water in Crisis: A Guide to the World’s Freshwater ResourcesOxford

University Press,” New York, 1993.

[5] “Lesson 1: The Water Crisis, Teacher Materials.” [Online]. Available:

http://nanosense.sri.com/activities/finefilters/watercrisis/FF_Lesson1Teacher.pdf.

[Accessed: 11-May-2016].

[6] “KSA water consumption rate twice the world average | Arab News.” [Online].

Available: http://www.arabnews.com/news/532571. [Accessed: 14-Apr-2016].

[7] J. Fawell, U. Lund, and B. Mintz, “Total dissolved solids in drinking-water.

Background document for development of WHO guidelines for drinking-water

quality,” World Health Organization, Geneva, 2003.

[8] N. Works, “Brackish groundwater: a viable community water supply option?,” The

National Water Commission, Australian …, 2011.

[9] A. E. Khan, A. Ireson, S. Kovats, S. K. Mojumder, A. Khusru, A. Rahman, and P.

Vineis, “Drinking Water Salinity and Maternal Health in Coastal Bangladesh:

Implications of Climate Change,” Environmental Health Perspectives, vol. 119,

no. 9, pp. 1328–1332, Apr. 2011.

[10] W. H. O. (WHO), “Guidelines for drinking-water quality 4th edition,” Geneva,

Switzerland, 2011.

[11] A. Cipollina, G. Micale, and L. Rizzuti, Seawater desalination: conventional and

renewable energy processes. 2009.

[12] O. Buros and I. D. Association, The ABCs of desalting. 2000.

[13] M. Abdel-Jawad, “Energy sources for coupling with desalination plants in the

GCC countries,” Consultancy report prepared for ESCWA,(September …, 2001.

[14] “Water Systems Aqua technology.” [Online]. Available:

http://www.aquatechnology.net/vaporcompressiondistillers.html. [Accessed: 05-

Apr-2015].

111

[15] “Reverse Osmosis units.” [Online]. Available: http://www.rotogo.co.uk/reverse-

osmosis.php. [Accessed: 05-Apr-2015].

[16] “Aqua Sky USA INC.” [Online]. Available:

http://www.aquaskyusa.com/under_sink_systems. [Accessed: 05-Apr-2015].

[17] J. Kucera, Reverse Osmosis: Design, Processes, and Applications for Engineers.

John Wiley & Sons, 2011.

[18] “Pressure-Driven Membrane Filtration Processes.” [Online]. Available:

http://synderfiltration.com/learning-center/articles/introduction-to-

membranes/pressure-driven-membrane-filtration-processes/. [Accessed: 11-May-

2016].

[19] “PURETEC Industrial Water - Article: What is Reverse Osmosis?” [Online].

Available: http://puretecwater.com/what-is-reverse-osmosis.html. [Accessed: 11-

May-2016].

[20] G. Osmonics, “The Filtration Spectrum,” GE Osmonics, Minnesota USA, 1996.

[21] M. A. Anderson, A. A. L. Cudero, and J. Palma, “Capacitive deionization as an

electrochemical means of saving energy and delivering clean water. Comparison to

present desalination practices: Will it compete?,” Electrochimica Acta, vol. 55, no.

12, pp. 3845–3856, Apr. 2010.

[22] S. A. W. R. C. electrodialysis AWWA Research Foundation, Lyonnaise des eaux-

Dumez (Firm), Water treatment membrane processes. New York: McGraw-Hill,

New York ©1996., 1996.

[23] A. Bernardes, M. Rodrigues, and J. Ferreira, “Electrodialysis and Water Reuse,”

Novel Approaches, Springer, 2013.

[24] A. Doyen, C. Roblet, L. Beaulieu, L. Saucier, Y. Pouliot, and L. Bazinet, “Impact

of water splitting phenomenon during electrodialysis with ultrafiltration

membranes on peptide selectivity and migration,” Journal of Membrane Science,

vol. 428, pp. 349–356, Feb. 2013.

[25] J. (Eds. . Bundschuh, J., & Hoinkis, Renewable Energy Applications for

Freshwater Production. CRC Press, 2012.

[26] U. Water, “The United Nations World Water Development Report 3–Water in a

Changing World,” 2009.

[27] F. Rahman and Z. Amjad, “14 Scale Formation and Control in Thermal

Desalination Systems, in Z. Amjad (Ed.),” in The Science and Technology of

Industrial Water Treatment, CRC Press, New York, 2010, pp. 271–296.

[28] R. Zhao, S. Porada, P. Biesheuvel, and A. Van der Wal, “Energy consumption in

membrane capacitive deionization for different water recoveries and flow rates,

and comparison with reverse osmosis,” Desalination, 2013.

112

[29] S. Porada, R. Zhao, A. van der Wal, V. Presser, and P. M. Biesheuvel, “Review on

the science and technology of water desalination by capacitive deionization,”

Progress in Materials Science, vol. 58, no. 8, pp. 1388–1442, Oct. 2013.

[30] M. Mossad and L. Zou, “A study of the capacitive deionisation performance under

various operational conditions,” Journal of hazardous materials, vol. 213–214, pp.

491–7, Apr. 2012.

[31] M. Mossad and L. Zou, “Effects of operational conditions on the electrosorption

efficiencies of Capacitive deionization,” Chemeca 2011: Engineering a Better

World: Sydney Hilton Hotel, NSW, Australia, 18-21 September 2011, p. 1648,

2011.

[32] S. Porada, L. Weinstein, R. Dash, A. van der Wal, M. Bryjak, Y. Gogotsi, and P.

M. Biesheuvel, “Water desalination using capacitive deionization with

microporous carbon electrodes.,” ACS applied materials & interfaces, vol. 4, no. 3,

pp. 1194–9, Mar. 2012.

[33] R. Zhao, S. Porada, P. M. Biesheuvel, and A. van der Wal, “Energy consumption

in membrane capacitive deionization for different water recoveries and flow rates,

and comparison with reverse osmosis,” Desalination, vol. 330, pp. 35–41, Dec.

2013.

[34] Y. Oren, “Capacitive deionization (CDI) for desalination and water treatment —

past, present and future (a review),” Desalination, vol. 228, no. 1–3, pp. 10–29,

Aug. 2008.

[35] A. M. Pernia, J. G. Norniella, J. A. Martin-Ramos, J. Diaz, and J. A. Martinez,

“Up–Down Converter for Energy Recovery in a CDI Desalination System,” IEEE

Transactions on Power Electronics, vol. 27, no. 7, pp. 3257–3265, Jul. 2012.

[36] M. Alkuran and M. Orabi, “Utilization of a buck boost converter and the method

of segmented capacitors in a CDI water purification system,” in 2008 12th

International Middle-East Power System Conference, 2008, pp. 470–474.

[37] A. M. Pernia, F. J. Alvarez-Gonzalez, M. A. J. Prieto, P. J. Villegas, and F. Nuno,

“New Control Strategy of an Up–Down Converter for Energy Recovery in a CDI

Desalination System,” IEEE Transactions on Power Electronics, vol. 29, no. 7, pp.

3573–3581, Jul. 2014.

[38] T. J. Welgemoed and C. F. Schutte, “Capacitive Deionization TechnologyTM: An

alternative desalination solution,” Desalination, vol. 183, no. 1–3, pp. 327–340,

Nov. 2005.

[39] M. E. Suss, S. Porada, X. Sun, P. M. Biesheuvel, J. Yoon, and V. Presser, “Water

desalination via capacitive deionization: what is it and what can we expect from

it?,” Energy Environ. Sci., vol. 8, no. 8, pp. 2296–2319, Jul. 2015.

[40] K. Laxmana, M. Myinta, and M. Al Abric, “Efficient Desalination of Brackish

113

Ground Water via a Novel Capacitive Deionization Cell Using Nanoporous

Activated Carbon Cloth Electrodes,” The Journal of Engineering Research

(TJER), Vol. 12, No. 2 (2015) 22-31.

[41] F. Ahmad, S. J. Khan, Y. Jamal, H. Kamran, A. Ahsan, M. Ahmad, and A. Khan,

“Desalination of brackish water using capacitive deionization (CDI) technology,”

Desalination and Water Treatment, vol. 57, no. 17, pp. 7659–7666, Apr. 2015.

[42] F. A. AlMarzooqi, A. A. Al Ghaferi, I. Saadat, and N. Hilal, “Application of

Capacitive Deionisation in water desalination: A review,” Desalination, vol. 342,

pp. 3–15, Jun. 2014.

[43] G.-H. Huang, T.-C. Chen, S.-F. Hsu, Y.-H. Huang, and S.-H. Chuang, “Capacitive

deionization (CDI) for removal of phosphate from aqueous solution,” Desalination

and Water Treatment, vol. 52, no. 4–6, pp. 759–765, Nov. 2013.

[44] L. Weinstein and R. Dash, “Capacitive Deionization: Challenges and

Opportunities,” Desalin. Water Reuse, vol. 23, pp. 34–37, 2013.

[45] R. Atlas and E. Aqua, “Purification of brackish water using hybrid CDI-EDI

technology,” presentation to International Desalination Conference, Aruba., 2007.

[46] P. M. Biesheuvel, R. Zhao, S. Porada, and A. van der Wal, “Theory of membrane

capacitive deionization including the effect of the electrode pore space.,” Journal

of colloid and interface science, vol. 360, no. 1, pp. 239–48, Aug. 2011.

[47] P. M. P. Biesheuvel, B. van Limpt, and A. van der Wal, “Dynamic

adsorption/desorption process model for capacitive deionization,” The Journal of

Physical Chemistry C, vol. 113, no. 14, pp. 5636–5640, Apr. 2009.

[48] P. M. Biesheuvel, S. Porada, M. Levi, and M. Z. Bazant, “Attractive forces in

microporous carbon electrodes for capacitive deionization,” Journal of Solid State

Electrochemistry, vol. 18, no. 5, pp. 1365–1376, May 2014.

[49] J. E. Dykstra, R. Zhao, P. M. Biesheuvel, and A. van der Wal, “Resistance

identification and rational process design in Capacitive Deionization.,” Water

research, vol. 88, pp. 358–70, Jan. 2016.

[50] K. Laxman, M. T. Z. Myint, M. Al Abri, P. Sathe, S. Dobretsov, and J. Dutta,

“Desalination and disinfection of inland brackish ground water in a capacitive

deionization cell using nanoporous activated carbon cloth electrodes,”

Desalination, vol. 362, pp. 126–132, Apr. 2015.

[51] S. Porada, B. B. Sales, H. V. M. Hamelers, and P. M. Biesheuvel, “Water

Desalination with Wires,” The Journal of Physical Chemistry Letters, vol. 3, no.

12, pp. 1613–1618, Jun. 2012.

[52] H. Helmholtz, “Ueber einige Gesetze der Vertheilung elektrischer Ströme in

körperlichen Leitern mit Anwendung auf die thierisch-elektrischen Versuche,”

Annalen der Physik und Chemie, vol. 165, no. 6, pp. 211–233, 1853.

114

[53] J. Araki, “Electrostatic or steric? – preparations and characterizations of well-

dispersed systems containing rod-like nanowhiskers of crystalline

polysaccharides,” Soft Matter, vol. 9, no. 16, p. 4125, 2013.

[54] M. Gouy, “Sur la constitution de la charge électrique à la surface d’un électrolyte,”

Journal de Physique Théorique et Appliquée, vol. 9, no. 1, pp. 457–468, 1910.

[55] D. L. Chapman, “LI. A contribution to the theory of electrocapillarity,”

Philosophical Magazine Series 6, vol. 25, no. 148, pp. 475–481, Apr. 1913.

[56] O. Stern, “Zur theorie der elektrolytischen doppelschicht,” Zeitschrift für

Elektrochemie und angewandte physikalische Chemie, vol. 30(21–22), pp. 508–

516., 1924.

[57] R. Zhao, P. M. Biesheuvel, and a. van der Wal, “Energy consumption and

constant current operation in membrane capacitive deionization,” Energy &

Environmental Science, vol. 5, no. Mcdi, p. 9520, 2012.

[58] P. M. Biesheuvel, R. Zhao, S. Porada, and a. van der Wal, “Theory of membrane

capacitive deionization including the effect of the electrode pore space,” Journal of

Colloid and Interface Science, vol. 360, pp. 239–248, 2011.

[59] T. Kim, J. E. Dykstra, S. Porada, A. van der Wal, J. Yoon, and P. M. Biesheuvel,

“Enhanced charge efficiency and reduced energy use in capacitive deionization by

increasing the discharge voltage.,” Journal of colloid and interface science, vol.

446, pp. 317–26, May 2015.

[60] R. Zhao, O. Satpradit, H. H. M. Rijnaarts, P. M. Biesheuvel, and a. van der Wal,

“Optimization of salt adsorption rate in membrane capacitive deionization,” Water

Research, vol. 47, no. 5, pp. 1941–1952, Apr. 2013.

[61] T. Humplik, J. Lee, S. C. O’Hern, B. A. Fellman, M. A. Baig, S. F. Hassan, M. A.

Atieh, F. Rahman, T. Laoui, R. Karnik, and E. N. Wang, “Nanostructured

materials for water desalination,” Nanotechnology, vol. 22, no. 29, p. 292001, Jul.

2011.

[62] T. J. Welgemoed and C. F. Schutte, “Capacitive Deionization TechnologyTM: An

alternative desalination solution,” Desalination, vol. 183, no. 1–3, pp. 327–340,

Nov. 2005.

[63] K. Laxman, M. T. Z. Myint, H. Bourdoucen, and J. Dutta, “Enhancement in ion

adsorption rate and desalination efficiency in a capacitive deionization cell through

improved electric field distribution using electrodes composed of activated carbon

cloth coated with zinc oxide nanorods.,” ACS applied materials & interfaces, vol.

6, no. 13, pp. 10113–20, Jul. 2014.

[64] J. Farmer, D. Fix, and G. Mack, “Capacitive deionization of NaCl and NaNO3

solutions with carbon aerogel electrodes,” Journal of the Electrochemical Society,

vol. 143, no. 1, p. 159, Jan. 1996.

115

[65] L. Pan, X. Wang, Y. Gao, Y. Zhang, Y. Chen, and Z. Sun, “Electrosorption of

anions with carbon nanotube and nanofibre composite film electrodes,”

Desalination, vol. 244, no. 1–3, pp. 139–143, Aug. 2009.

[66] L. Zou, L. Li, H. Song, and G. Morris, “Using mesoporous carbon electrodes for

brackish water desalination.,” Water research, vol. 42, no. 8–9, pp. 2340–8, Apr.

2008.

[67] A. Aghigh, V. Alizadeh, H. Y. Wong, M. S. Islam, N. Amin, and M. Zaman,

“Recent advances in utilization of graphene for filtration and desalination of water:

A review,” Desalination, vol. 365, pp. 389–397, Jun. 2015.

[68] C. Nie, Y. Zhan, L. Pan, H. Li, and Z. Sun, “Electrosorption of different cations

and anions with membrane capacitive deionization based on carbon

nanotube/nanofiber electrodes and ion-exchange membranes,” Desalination and

Water Treatment, vol. 30, no. 1–3, pp. 266–271, Aug. 2012.

[69] W. Huang, Y. Zhang, S. Bao, and S. Song, “Desalination by capacitive

deionization with carbon-based materials as electrode: a review,” Surface Review

and Letters, 2013.

[70] J. Lee, W. Bae, and J. Choi, “Electrode reactions and adsorption/desorption

performance related to the applied potential in a capacitive deionization process,”

Desalination, 2010.

[71] J.-H. Lee and J.-H. Choi, “The production of ultrapure water by membrane

capacitive deionization (MCDI) technology,” Journal of Membrane Science, vol.

409–410, pp. 251–256, Aug. 2012.

[72] A. (2014). El Shahat, “Empirical Capacitive Deionization ANN Nonparametric

Modeling for Desalination Purpose.,” Journal of Engineering Research and

Technology, vol. 1, no. 2, pp. 58–65, 2014.

[73] M. Mossad, W. Zhang, and L. Zou, “Using capacitive deionisation for inland

brackish groundwater desalination in a remote location,” Desalination, vol. 308,

pp. 154–160, Jan. 2013.

[74] J. C. Farmer, D. V. Fix, G. V. Mack, R. W. Pekala, and J. F. Poco, “Capacitive

deionization of NH4ClO4 solutions with carbon aerogel electrodes,” Journal of

Applied Electrochemistry, vol. 26, no. 10, pp. 1007–1018, 1996.

[75] J. Farmer, “Method and apparatus for capacitive deionization, electrochemical

purification, and regeneration of electrodes,” US Patent 5,425,858, 1995.

[76] K. Dermentzis, “Continuous electrodeionization through electrostatic shielding,”

Electrochimica Acta, 2008.

[77] K. Dermentzis and D. Papadopoulou, “A new process for desalination and

electrodeionization of water by means of electrostatic shielding zones-ionic current

sinks,” Journal of Engineering Science and Technology Review, no. 2(1), pp. 33–

116

42., 2009.

[78] J. H. Ryu, T. J. Kim, T. Y. Lee, and I. B. Lee, “A study on modeling and

simulation of capacitive deionization process for wastewater treatment,” Journal of

the Taiwan Institute of Chemical Engineers, vol. 41, no. 4, pp. 506–511, 2010.

[79] Y. a C. Jande and W. S. Kim, “Modeling the capacitive deionization batch mode

operation for desalination,” Journal of Industrial and Engineering Chemistry, vol.

20, no. 5, pp. 3356–3360, 2014.

[80] P. M. Biesheuvel and a. van der Wal, “Membrane capacitive deionization,”

Journal of Membrane Science, vol. 346, pp. 256–262, 2010.

[81] S. Porada, L. Weinstein, R. Dash, a. Van Der Wal, M. Bryjak, Y. Gogotsi, and P.

M. Biesheuvel, “Water desalination using capacitive deionization with

microporous carbon electrodes,” ACS Applied Materials and Interfaces, vol. 4, pp.

1194–1199, 2012.

[82] R. Zhao, O. Satpradit, H. H. M. Rijnaarts, P. M. Biesheuvel, and A. van der Wal,

“Optimization of salt adsorption rate in membrane capacitive deionization.,” Water

research, vol. 47, no. 5, pp. 1941–52, Apr. 2013.

[83] H. Ni, R. C. Amme, and Y. Jin, “Desalination by electret technology,”

Desalination, vol. 174, pp. 237–245, 2005.

[84] R. Zhao, P. M. Biesheuvel, H. Miedema, H. Bruning, and A. van der Wal, “Charge

Efficiency: A Functional Tool to Probe the Double-Layer Structure Inside of

Porous Electrodes and Application in the Modeling of Capacitive Deionization,”

The Journal of Physical Chemistry Letters, vol. 1, no. 1, pp. 205–210, Jan. 2010.

[85] S. Jeon, H. Park, J. Yeo, S. Yang, C. H. Cho, M. H. Han, and D. K. Kim,

“Desalination via a new membrane capacitive deionization process utilizing flow-

electrodes,” Energy & Environmental Science, vol. 6, no. 5, p. 1471, 2013.

[86] Y. Gendel, A. K. E. Rommerskirchen, O. David, and M. Wessling, “Batch mode

and continuous desalination of water using flowing carbon deionization (FCDI)

technology,” Electrochemistry Communications, vol. 46, pp. 152–156, Sep. 2014.

[87] A. Rommerskirchen, Y. Gendel, and M. Wessling, “Single module flow-electrode

capacitive deionization for continuous water desalination,” Electrochemistry

Communications, vol. 60, pp. 34–37, Nov. 2015.

[88] S. Porada, L. Borchardt, M. Oschatz, M. Bryjak, J. S. Atchison, K. J. Keesman, S.

Kaskel, P. M. Biesheuvel, and V. Presser, “Direct prediction of the desalination

performance of porous carbon electrodes for capacitive deionization,” Energy &

Environmental Science, vol. 6, no. 12, p. 3700, Nov. 2013.

[89] Y. A. C. Jande and W. S. Kim, “Modeling the capacitive deionization batch mode

operation for desalination,” Journal of Industrial and Engineering Chemistry, vol.

20, no. 5, pp. 3356–3360, Sep. 2014.

117

[90] A. Omosebi, X. Gao, J. Landon, and K. Liu, “Membrane Assisted Capacitive

Deionization with Binder-Free Carbon Xerogel Electrodes,” ECS Transactions,

vol. 61, no. 21, pp. 9–17, Oct. 2014.

[91] K. Laxman, L. Al Gharibi, and J. Dutta, “Capacitive deionization with asymmetric

electrodes: Electrode capacitance vs electrode surface area,” Electrochimica Acta,

vol. 176, pp. 420–425, Sep. 2015.

[92] D. MacGregor and M. Smith, “Method and apparatus for separating ions from

liquids to produce separate diluted and concentrated effluents,” US Patent

4,948,514, 1990.

[93] D. MacGregor, “Method for separating ions from liquids,” US Patent 5,061,376,

1991.

[94] “Aqua EWP.” [Online]. Available: http://aquaewp.com/. [Accessed: 11-May-

2015].

[95] “Plimmer – CDI.” [Online]. Available: http://www.idropan.com/en/portfolio-

view/plimmer-cdi/. [Accessed: 11-May-2015].

[96] “Voltea - Unlocking the world’s water potential.” [Online]. Available:

http://www.voltea.com/#1. [Accessed: 11-May-2015].

[97] “Atlantis Technologies.” [Online]. Available: http://www.atlantis-

water.com/about-us. [Accessed: 11-May-2015].

[98] M. Hayashi, “Temperature-Electrical Conductivity Relation of Water for

Environmental Monitoring and Geophysical Data Inversion,” Environmental

Monitoring and Assessment, vol. 96, no. 1–3, pp. 119–128, Aug. 2004.

[99] P. N. Sen and P. A. Goode, “Influence of temperature on electrical conductivity on

shaly sands,” GEOPHYSICS, vol. 57, no. 1, pp. 89–96, Jan. 1992.

[100] “ZORFLEX ACC Activated Knitted Carbon Cloth datasheet.” [Online]. Available:

http://www.calgoncarbon.com/wp-content/uploads/product-

literature/Zorflex_Knitted.pdf. [Accessed: 13-Apr-2016].

[101] “3M XYZ-Axis Electrically Conductive Tape 9713 datasheet.” [Online].

Available: http://multimedia.3m.com/mws/media/66127O/3mtm-xyz-axis-

electrically-conductive-tape-9713.pdf. [Accessed: 16-May-2016].

118

Vitae

Name : Syed Adnan Ali

Nationality : Indian

Date of Birth : 9/8/1992

Email : syedadnanali26@gmail.com

Address : 16-9-438/1, Old Malakpet, Telangana, Hyderabad, 500036

Academic Background : B.E. in Electronics and Communication Engineering,

2013, Osmania University, Hyderabad, India.

Conferences Attended : Moustafa Elshafei, Syed Adnan Ali, “Modelling and

Control of Travelling wave Capacitive Deionization

Desalination System", CDI&E 2015, International

Conference on Capacitive Deionization and

Electrosorption, Saarbrücken, Germany, October 26-29,

2015.

Moustafa Elshafei, Syed Adnan Ali, “Modeling of

capacitive de-ionization with activated carbon

electrodes”, 2nd International Conference on Desalination

and Environment (ICODE), Doha, Qatar, January 23-26,

2016.

Journal publications : Modeling of capacitive de-ionization with activated

carbon electrodes (under review), by Desalination and

Water Treatment Journal.